From larry.sanger at corp.manyone.net Mon May 15 23:22:00 2006 From: larry.sanger at corp.manyone.net (Larry Sanger) Date: Mon, 15 May 2006 23:22:00 -0700 Subject: [Textop] Test textop Message-ID: <004901c678b1$0a8969c0$a0f5fea9@D7WRQ591> Test textop From blarneypilgrim at yahoo.com Mon May 15 23:22:18 2006 From: blarneypilgrim at yahoo.com (Larry Sanger) Date: Mon, 15 May 2006 23:22:18 -0700 Subject: [Textop] Test textop Message-ID: <004a01c678b1$157d3730$a0f5fea9@D7WRQ591> Test textop From larry.sanger at corp.manyone.net Wed May 17 15:07:03 2006 From: larry.sanger at corp.manyone.net (Larry Sanger) Date: Wed, 17 May 2006 15:07:03 -0700 Subject: [Textop] Outlining as a matter of taste and as a management problem In-Reply-To: Message-ID: <000001c679fe$3a75ca90$a0f5fea9@D7WRQ591> All, The last week has had us very busy with a new launch--you can see the new DUF website at http://www.dufoundation.org and either today or tomorrow you should be able to see the DU itself *without the M1 browser*, at http://www.digitaluniverse.net --but as of this writing, the latter URL is still the old website. I've announced Textop and a lot of other things (DU Journal, mailing lists) on my shiny new *real* blog http://www.dufoundation.org/blog/, real because it's not just static HTML pages but actually runs blogging software. :-) I'll probably be blogging daily or almost daily, that's the hope. But since no one yet knows about dufoundation.org, no one knows about the blog... Anyway, soon (tomorrow?) I'll be asking you to spread the word about Textop. Responding now finally to Kunal Sen. I have to say it's a privilege to be able to discuss these matters with a technical ontologist on the one hand, and a person who worked on the taxonomy for the Britannica on the other. Kunal raises two issues and both of them are well worth discussing: (1) Isn't outlining inherently subjective and arbitrary? > However, I'd like to comment on one of his early observations > -- "informal hierarchies of topics are very much arbitrary > (there is no "right place" for a node, placing nodes is a > matter a personal preferences/goals)". I have been involved > with the creation of the taxonomic structure for a large > general reference encyclopedia. Here a small group of people > were responsible for creating the structure, and they all > knew each other, had definite subject areas that each > individual was responsible for, and there was centralized > editorial control, and yet the location of a particular node > often reflected their personal biases. In some cases it came > down to matters of taste. Many people have told me that they see a problem in the fact that, when one is sorting concepts into an outline, frequently the same concept needs to live in multiple locations, and there is no reasoned way to choose one structure over another--as if I were not a philosopher and did not already know this deep in my bones. In fact, this is a point that the philosopher R. G. Collingwood spent what seemed to me a needless amount of space proving in *An Essay on Philosophical Method*--needless, because it's pretty obvious. What I am interested in exploring, however, is the possibility that *very specific* propositions, arguments, explanations, etc., *are* amenable to *fewer* organizations. One main reason that it is difficult to come up with a perfectly coherent set of academic department titles, for example, or library cataloging system, is that the things thus organized and distinguished (academic subjects and books) are such messy things. They overlap in their subject matter all the time. But that is *not* the case, or not so obviously, when we're talking about very specific, paragraph-sized chunks of text. If a text is well-written, it concerns just one or two things. It might *mention* a zillion topics, but makes just one or two *main*, interesting points, that the author is specifically trying to convey. What I think I discovered in writing the outline of Hobbes' *Leviathan* is that, when the items to be organized are that specific, there is much less overlap, and it becomes remarkably easier than one might expect--by no means *easy*, but easier than expected--to create an outline that structures the definitions, propositions, arguments, etc. contained in the text. Moreover, I found that the more that I worked through the *Leviathan*, the more better I thought the outline was becoming. Extra "data" in the form of chunks tended to confirm or disconfirm certain classificatory choices I made. I have no doubt whatsoever that I could continue executing my personal plan--which called for going through Locke's *Essay* next and putting it into the same outline--and the outline would become better in every respect. I look forward to collating certain works precisely because they will help form and solidify areas of the outline that are now very crude. As I collect more and more chunks under a given node, I make children nodes, or place some of the texts in a different branch, or whatever. There will always be some redundancy in placement of chunks (I not infrequently put the same chunk in two different places) and practical arbitrariness (choosing a new header wording to cover a text on some new topic frequently seemed ad hoc). What I'm suggesting, however, is that outlining will be easier, clearer, and on a decided vector of improvement precisely because we're working with the *details* of *enormous numbers* of texts--not with relatively vague or broad things, like concepts, books, academic departments, encyclopedia articles, or many of the other things one ordinarily speaks of classifying. (2) How should we manage decisionmaking about the outline? > Therefore, when the same is done by a large unconnected group > of contributors, at a much grander scale, subjective biases > are bound to enter the structure. So the question is, how can > they be minimized, is such a structure viable, and how such > biases would affect the effective use of such a system? I > think this is the time to take a hard look at these questions > and make sure we are on the right path. A collaborative > project like this is hard to steer once it is launched. I agree 100% with what Kunal says here. How can we reach agreement, a working consensus, on the shape of the outline? I agree with Philippe Martin that we absolutely need clear rules--not necessarily *very* rigorous rules, but rules clear enough to be permit ordinary educated people to make and defend decisions, and simple enough for such people to understand. But I also do not think that we can know what the rules are, or how best to enforce them and resolve disagreements, until we have had some experience actually playing with a system *very* like the one we're proposing to set up. That's why I think we need to do a pilot project (http://www.textop.org/pilot_project.html). I'm next going to post a *draft* set of rules, which I'm not even comfortable posting on the website yet, for discussion and as an example. --Larry From larry.sanger at corp.manyone.net Wed May 17 18:09:49 2006 From: larry.sanger at corp.manyone.net (Larry Sanger) Date: Wed, 17 May 2006 18:09:49 -0700 Subject: [Textop] Collation Project rules (very rough draft) Message-ID: <000501c67a17$c2bbc580$a0f5fea9@D7WRQ591> By request, here is a very rough draft of rules for the Collation Project: http://www.textop.org/collation_principles.html Not yet linked from the main site, because I need to think more about them. Feedback about any or all of them would be nice. --Larry From phmartin at phmartin.info Fri May 19 08:21:27 2006 From: phmartin at phmartin.info (Philippe MARTIN) Date: Sat, 20 May 2006 01:21:27 +1000 Subject: [Textop] FW: still not fine-grained and structured enough to be scalable Message-ID: <200605191521.k4JFLRap011429@pc070372.sci.griffith.edu.au> Dear all, Below are some notes/answers I wrote on Saturday May 13th about the comments/questions of Kunal and Larry regarding my original message. I could not send these notes earlier because the mailing list previous handler had problems with my email address. Although these comments are still valid I am not sure that they are important anymore for this project. Indeed, in light of the recent information I agree with Larry's recent remark (in a direct email) that I am trying to achieve different goals, even though I also agree with his descriptions of an "Ideal Information Resource" in his article titled "The Future of Free Information" (Digital Universe Journal, 2006-1). Although my implementation/project sometimes better fits the descriptions of this article, I agree with Larry that his implementation/project is more adequate for achieving the sub-goal of "building user-friendly reference works for largely non-technical people". So I think these two implementations/approaches are complementary and that they should point to each other, to permit people who want to find and update more (vs. less) fine grained info units than paragraphs and topics (vs. sentences and concepts) to find an adequate site. These people may even participate to both (but it is clear that most persons will prefer Larry's approach). There are also a lot of possible cross-overs between the two approaches: I allow paragraphs, and the relations I propose could be re-used in the TextOp or Debate projects. So I consider my work as part of the TextOp Pilot project and hence would like you to consider the next 2 documents: http://www.webkb.org/kb/classif/sdLeviathan.html http://www.webkb.org/kb/classif/sdGodExistence.html The first represents/indexes statements and concepts from the first five chapters of the Leviathan, and the second represents and complements the argumentation tree used in John Locke's Chapter 9 of his essay "Of our Threefold Knowledge of Existence". The interface to access and browse those representations from the lexical ontology of WebKB-2 and from related representations will be available in July (it will basically be the current interface of WebKB-2 but with more presentation options). Philippe ============= My notes/answers of Saturday May 13th ================== > If I understood correctly, he believes that in order to be useful > and sustainable the system must be created with a formal logical > structure, grounded in theoretical work done in the area of semantic > knowledge representations, instead of the current intuitive structure. "semi-formal logic-based structure" (indeed, very few human beings could really bear a formal logic-based structure for any realistic amount of any realistically complex kind of knowledge). Some people find semi-formal logic-based structures intuitive, other do not (at first). My approach does not require people to be logicians but I concede that it might require them to be "severely analytically-minded" and I do agree that this would limit the number of participants. > Philippe, I apologize for this long and vigorous reply, but I hope you > will take that as a sign of my respect for and interest in your ideas. I do. Thanks for your long and vigorous reply. In fact, even though I carefully read your example and more quickly read all the other pages I have probably over-interpreted your project and hence my message was the one that was a bit too vigorous. This may still be the case below. But as you wrote, it was and is a way to be helpful by raising issues and elicit precisions about the project. > As Kunal wisely said, to do this justice we'd have to read your > references. So, perhaps I'm not understanding at all, but this is not > really sounding very much like what I'm proposing to do. Indeed. It is difficult to understand what I mean without reading the first part of my ICCS'05 article (http://www.webkb.org/doc/papers/iccs05/), looking at examples such as those accessible from http://www.webkb.org/kb/classif/sd.html#examples AND playing a bit with WebKB (http://www.webkb.org). I also still have problems in understanding your approach, as indicated below. > But the outline of the Collation Project *is not* an ontology, nor is > it meant to be one. Again, consult the example. ... > Of course there will be thousands of paragraphs, if the > "idea/topic/statement" is broad enough. Your example is about one source document. I have no problem with it although I'd personally be much more interested in a detailed argumentation graph such as in my above cited examples. My problem is that I fail to imagine how your mono-source example generalise into one structure that organises the topics and arguments of thousands of documents, without using an ontology to solve the problem of information organisation and retrieval. I certainly agree that there will be a lot of redundancies in the sources that people will exploit and hence they can for example choose not to represent arguments that are equivalent to arguments that have previously been represented. However, for each idea/statement there will still be a *lot* of arguments, sub-arguments, counter-arguments on the sub-arguments, etc. How will they be organised (and then compared) if specialization and argumentation relations are not used? It may be that you do not want the sub-arguments to be organised ("the items I propose to classify are these sorts of 'text chunks'") but then (i) information retrieval and comparison may well be an issue, (ii) it would clearly be "of more interest" to organize them (if people are willing to do that) even though your approach is already very interesting. Furthermore, regarding "topics" now, how the tens or hundreds of thousands of topics that the text chunks are about will be organized (and hence retrieved) if specialization and other relations are not explicitely used between these concepts/topics? A Yahoo/DMOZ like hierarchy of topics would be hopeless (if we do not agree on that, please tell me but I believe we do). Keyword-based retrieval would be hopeless too. You are talking about relations in your answer ("...designate what the parent-child relations shall mean... more definite relations ... in relation to a cluster of other related concepts ... what relation a specific point about realism ...) but not what kinds of relations nor if they would always be explicited. If there are different kinds of relations (and there would need to be to avoid a Yahoo/ODP like hierarchy of topics), the relations will need to be explicit and your hierarchy of topics will actually be a semantic network and an ontology (even if it happens to be only a semi-formal ontology). I assume that thousands of your top-level topics will use English words or expressions. Then, a lexical ontology of English is useful because it already organises and permits to retrieve those topics and guide the placement of new topics (you can test this by entering an English noun for a topic at http://www.webkb.org/). In fact, at this point, these are not just topics, these are concepts. (And the ontology needs to be updatable because it does not include all the necessary concepts). > And how we can expect people constructing the outline just to > buy into the ontology wholesale? Please tell me/us what kinds of relations you envisage between the text chunks and I think I will undertand your approach. Again, if the relations are explicit and have a known meaning, this is an ontology (strictly speaking, even a list of terms is an ontology but I am raising the bar a little bit). If your own experience tells you that explicit relations are not needed for your project, I do not think there is a way I can convince you. For what you have in mind, explicit relations may actually not be needed. > So I guess I need to know is how this network of statements is > to be generated: by summarizing texts, for example? > Or by producing them ourselves, *a priori* as it were, and then > expecting texts to fall neatly into the network? A "good" argumentation graph cannot be generated automatically (I am not advocating mere "concept maps") since many argumentation steps in the sources are either implicit or simetimes bogus. Creating a good argumentation graph is difficult even for people and it forces them to be careful in their statements. Although some research projects work on the extraction of rhetorical relations or argumentation relations from documents by using NLP techniques, I am not sure this would help to create good argumentation graphs. I do not see any "a priori" in representing arguments, except in the act of interpreting/understanding the source text (assuming that the interpreter is not the creator of the source text). I do not understand the sentence "expecting texts to fall neatly into the network" since the texts will not change and the network is not predefined but extracted from the texts. > Are you saying that the project should *rate* particular > statements (outline nodes, I guess) as "correct" or "interesting"? Not the project but each user of the project if s/he wishes to (although project members may be given much more weight), and not "topics" but any "assertion/statement" that the user wishes to rate With my approach, this also applies to any relation between topics or statements, and this is intended to ease the personalization of information retrieval/filtering and permit "democratic" cooperation (or in other words, minimising the need for commitees to make choices). I am afraid that the sections 2.1 and 2.2 of my article (http://www.webkb.org/doc/papers/iccs05/) really need to be read to understand this. > What would the input interface be like? > How would the *result* look different from what I've illustrated? A good part of the interface/results would look like the current interface/results of WebKB-2 (see also the above cited examples) although with more options to filter out unwanted details, with more understandable notations, and much better menus to guide the adding of relations. If you think this would be too frightening for the users and do not want to take that path, I understand. I guess it is a matter of preferences; from my viewpoint, I do not see a better path. > How would *the collation of texts* proceed? > What's the overall procedure? My answer to this kind of question is generally: choose any methodology you wish or none at all, just make sure you respect (or make the users respect) the semantics of the relations they use, this will permit an incremental and correct building of the knowledge base (I also assume cooperation protocols such as those I propose if there are many knowledge providers). This certainly worked for me each time I represented knowledge and, although I acknowledge that knowledge acquisition methodologies are interesting, I have not seen any ontology engineering methodology that really was doing anything else that providing a set of top-level concepts and relations, and giving advices on how to use them. Surprisingly (for me), so far the above answer rarely satisfied the persons who asked that kind of question. I doubt this will be the case this time but I still hope to. However, I can provide the merged set of the top-level concepts and relations advocated by many methodologies, including argumentation methodologies. For organising statements, I found the following relations useful: specialisation, correction, corrective_specialization, more_concise_reformulation, more_precise_reformulation, objection, argument, example, opposition and consistent_with. Philippe _______________________________________________________________________ Dr. Philippe Martin Address: Griffith Uni, School of ICT, PMB 50 GCMC, QLD 9726 Australia Web site: www.phmartin.info; Tel: +61 7 5552 8252; Fax: +61 7 5552 8066 _______________________________________________________________________ From larry.sanger at corp.manyone.net Fri May 19 12:08:02 2006 From: larry.sanger at corp.manyone.net (Larry Sanger) Date: Fri, 19 May 2006 12:08:02 -0700 Subject: [Textop] New arrival Message-ID: <006e01c67b77$8d2dee00$a0f5fea9@D7WRQ591> All, I apologize for my silence and I'll have to apologize in advance for continued silence until Monday at least. My first child, a boy, arrived last night! --Larry From webmaster at gblane.com Fri May 19 13:49:34 2006 From: webmaster at gblane.com (gblane) Date: Fri, 19 May 2006 13:49:34 -0700 (PDT) Subject: [Textop] New arrival In-Reply-To: <006e01c67b77$8d2dee00$a0f5fea9@D7WRQ591> Message-ID: <20060519204934.39507.qmail@web30206.mail.mud.yahoo.com> Congratulations and best wishes to your family. GB Larry Sanger wrote: All, I apologize for my silence and I'll have to apologize in advance for continued silence until Monday at least. My first child, a boy, arrived last night! --Larry _______________________________________________ Textop mailing list Textop at lists.dufoundation.org http://lists.dufoundation.org/mailman/listinfo/textop -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.dufoundation.org/archive/textop/attachments/20060519/5d94cbfb/attachment.html From esautede at riccimac.org Sun May 21 04:28:33 2006 From: esautede at riccimac.org (Eric Sautede) Date: Sun, 21 May 2006 19:28:33 +0800 Subject: [Textop] New arrival In-Reply-To: <006e01c67b77$8d2dee00$a0f5fea9@D7WRQ591> Message-ID: Congrats! Eric ______________________________________________________ Eric Sautede Research fellow and Chief Editor of Chinese Cross Currents > From: Larry Sanger > Reply-To: General Discussion List for the Text Outline Project > > Date: Fri, 19 May 2006 12:08:02 -0700 > To: > Subject: [Textop] New arrival > > All, > > I apologize for my silence and I'll have to apologize in advance for > continued silence until Monday at least. My first child, a boy, arrived > last night! > > --Larry > > _______________________________________________ > Textop mailing list > Textop at lists.dufoundation.org > http://lists.dufoundation.org/mailman/listinfo/textop > From mpoe at theatlantic.com Mon May 22 04:48:20 2006 From: mpoe at theatlantic.com (mpoe) Date: Mon, 22 May 2006 07:48:20 -0400 Subject: [Textop] New arrival In-Reply-To: Message-ID: Great news, Larry. My heartfelt congratulations to you and your growing family! Best, MP Marshall Poe, Ph.D. Editor, MemoryArchive www.memoryarchive.org mpoe at theatlantic.com On 5/21/06 7:28 AM, "Eric Sautede" wrote: > Congrats! > Eric > ______________________________________________________ > Eric Sautede > Research fellow and Chief Editor of Chinese Cross Currents > > > >> From: Larry Sanger >> Reply-To: General Discussion List for the Text Outline Project >> >> Date: Fri, 19 May 2006 12:08:02 -0700 >> To: >> Subject: [Textop] New arrival >> >> All, >> >> I apologize for my silence and I'll have to apologize in advance for >> continued silence until Monday at least. My first child, a boy, arrived >> last night! >> >> --Larry >> >> _______________________________________________ >> Textop mailing list >> Textop at lists.dufoundation.org >> http://lists.dufoundation.org/mailman/listinfo/textop >> > > > _______________________________________________ > Textop mailing list > Textop at lists.dufoundation.org > http://lists.dufoundation.org/mailman/listinfo/textop From EXT-Mathias.Brendel at nokia.com Tue May 23 01:14:14 2006 From: EXT-Mathias.Brendel at nokia.com (EXT-Mathias.Brendel at nokia.com) Date: Tue, 23 May 2006 11:14:14 +0300 Subject: [Textop] Governing system Message-ID: <5D2D795A0116244CABFE285440AA48AA0175EEFC@ouebe101.NOE.Nokia.com> Hi all, Larry congratulations. Responding to the questions of governance. 1) In some of the questions, we agree that "lets see", or we "agree to disagree", but it is best to see the pilot project. I do not respond to these parts. In some other questions I would like to put here some arguments, emphasize my point of view. 3) >the issue that Kunal Sen raised: "A collaborative project like >this is hard to steer once it is launched." I agree with this. I just go back to this, because this is new, this is a meta-question and it is important. So it is important to construct the governance system appropriately. 4) >Only then will we be able to say with any certainty what role >editors ought to play. For me it is reasonable at present that there shall be at least 2 roles beside the collaborator. One is the editor. He has power to reject a piece of work to be published. But he has no 5) >I see--and I would agree, so long as we are talking about >elections among the Editorial Committees. My fear along these lines is >*another* problem with Wikipedia: early on, less well-qualified people >flooded the project and began shouting down those with specialized >knowledge, driving them off. I agree, this is a big problem. We must take the advantage to organize the system in a good way, especially in the governance system. In my opinion, if there is a policy formulated, which states explicitly that for example editors are expected to have PhD in the area, and we have a well established group of editors, then it is hard for the "mob" to make a revolution, unless they are right. The "mob" does not organize itself against a well established system, until there is not a very strong reason. So I think that a set of good editors, with the help of well written policies should gain the respect of the majority in a democratic system. 6) >power. Then I have another question for you: how is it >determined who is in "the rank and file" (who is a basic >member) of the project? I would say the solution in Wikipedia was good enough. Collaborators, who spent some time and had a sufficient amount of contribution, may vote. 7) >I see. The American term for it is the separation of powers, >and again I agree in principle. But as always, I need >specifics: what functions are you conceiving as "administrative" >functions, what functions are "editorial," >and why (in view of what danger) should those functions not be shared >by the same people? Yes, separation of power. An editor is somebody, who has control over publishing something. But he shall not have power over people. The administrator can ban somebody (details to be worked out), but shall not have power over editorial questions. The danger I see here is as follows: lets say a collaborator argues on some issue with the editor. They might go into a quarrel about the issue. If the editor would have administrative power, then his judgement can be biased. My experience in Wikipedia was that administrators were getting a special power, since they could banned everybody. This meant after a while, that all contentual issues were decided eventually by the authority of the administrator. 8) >Well, if we're talking about a self-serving dictatorship of a >fundamentally undemocratic nation, then of course I'm all in favor of >control from down below. But we're talking now about Textop, and the >goals of its governance are (I think) to make excellent decisions >quickly, to retain as many valuable people as possible, etc. Why think >that using *all* members to elect editors will help to serve such >functions? I do not insist on all members. What I insist on, that it must be regularly and pre-defined. In regular time-spans there must be a real check of confidence. 9) >Again: why? Just because "it's more democratic"? If we agree that a >distributed content project is not precisely the same as a genuine >real-world polity, then you need to say more than "it's more >democratic" to defend a policy stand. More democratic system means more objective, or more NPOV content in the end. 10) One issue in addition. Shall we create a voting system, which is anonymous? There are arguments for this and against this. A) For anonymity is that nobody else sees, what I vote for, or who I vote for. We know, why this is good in real life, this is also good in such a system B) But in real life, there is a great confidence in the voting system. In such a project somebody may be suspicious, that somebody is rigging the voting. 10) In Wikipedia the system of policies was very poorly constructed. In the Hungarian pages there was no real page collecting all them togather. One could not find a good starting point. It was not clear, what are accepted policies, what is the formal way of accepting policies, is there a formal way at all? It would be desired that policies are structured in a good system, with a good starting point and a good hierarchy. Mathias From EXT-Mathias.Brendel at nokia.com Tue May 23 23:30:44 2006 From: EXT-Mathias.Brendel at nokia.com (EXT-Mathias.Brendel at nokia.com) Date: Wed, 24 May 2006 09:30:44 +0300 Subject: [Textop] Role-play-game In-Reply-To: <5D2D795A0116244CABFE285440AA48AA0175EEFC@ouebe101.NOE.Nokia.com> Message-ID: <5D2D795A0116244CABFE285440AA48AA0179D577@ouebe101.NOE.Nokia.com> Hi, I noticed that I had a point unfinished. This is because I noticed that I will explain it later. But I did not delete the unfinished part. So "For me it is reasonable at present that there shall be at least 2 roles beside the collaborator. One is the editor. He has power to reject a piece of work to be published. But he has no" The explanation fotr this was this: "An editor is somebody, who has control over publishing something. But he shall not have power over people. The administrator can ban somebody (details to be worked out), but shall not have power over editorial questions." So these are 2 important roles in my opinion. There is sure a role of a general collaborator. Further, there could be a role of arbitrator. We did not have this in the Hungarian Wikipedia, and I think, this is important. If two parties are quareling, it is important to have a third point of view. The arbitrator shall be independent. Here again, separation of roles is important. There is no use of an arbitrator, if he is also an administrator. Checkuser: this was a role in Wikipedia, which I do not think is good. Actually the problem is not with the role, but with the question of anonymity. I think, we shall not allow anonym collaborators. Moreover, I think it would be desired that every collaborator takes part with his real name. We shall know him/her. That means that we shall know his/her background. I have no detailed idea about this. I can imagine many solution. Shall we require some kind of a CV? In any case, a collaborator with a good background shall get high respect in the project. Matthias From larry.sanger at corp.manyone.net Thu May 25 20:46:52 2006 From: larry.sanger at corp.manyone.net (Larry Sanger) Date: Thu, 25 May 2006 20:46:52 -0700 Subject: [Textop] The Digital Universe launches a bunch of stuff Message-ID: <002601c68077$066a9670$a0f5fea9@D7WRQ591> All, I've summarized what we've launched here: http://www.dufoundation.org/blog/ Seven items! Eight if you include my firstborn child. I hope to start catching up more seriously with Textop stuff tomorrow...in the meantime, I am hoping there might be further discussion of the http://www.textop.org/pilot_project.html If there's no further discussion I might just go ahead and install MediaWiki in the Textop Web space and try to get things going! --Larry From mdflood at starpower.net Fri May 26 06:26:11 2006 From: mdflood at starpower.net (M Flood) Date: Fri, 26 May 2006 09:26:11 -0400 Subject: [Textop] Signing up Message-ID: <447701F3.5010106@starpower.net> Greetings all: I have a simple administrative question that may be of interest to others. (I hope I'm not the only one who can't figure this out!) I signed up several months ago to learn more about the stewardship program. Now that there's been an official call for participation, I looked at the new recruitment page (www.dufoundation.org/stewards.php), which has a sign-up form that appears slightly different from what I used before. Should I sign up again, or am I already in the database? Thanks -- Mark Flood From larry.sanger at corp.manyone.net Sat May 27 19:46:27 2006 From: larry.sanger at corp.manyone.net (Larry Sanger) Date: Sat, 27 May 2006 19:46:27 -0700 Subject: [Textop] Textop wiki set up Message-ID: <000001c68200$eae22d60$a0f5fea9@D7WRQ591> http://www.textop.org/wiki All, I've set up MediaWiki in the Textop space; see above. I've put up the pilot project proposal on the main page, ready for you to edit (or comment on, here: http://www.textop.org/wiki/index.php?title=Talk:Main_Page&action=edit). I'll be getting caught up on textop mailing list mail soon...in the meantime, the question foremost in my mind now is whether we ought to try out a pilot project soon, as outlined above, or whether we should (for some reason) postpone the pilot project. --Larry From larry.sanger at corp.manyone.net Mon May 29 18:33:23 2006 From: larry.sanger at corp.manyone.net (Larry Sanger) Date: Mon, 29 May 2006 18:33:23 -0700 Subject: [Textop] [textop] Common assumptions; governance. Message-ID: <001801c68389$0a84c770$a0f5fea9@D7WRQ591> All, Well, baby Henry is now 12 days old, and papa is getting a little extra time to get other things done (courtesy grandma). So I am going to try to catch up with some Textop mail. I've been very happy with all the discussion so far. Thank you all, by the way, for the congratulations on Henry's arrival. Responding to Matthew Elvey here. > >>> Administrators can not be [editors] and vice versa. > > I like this idea; the separation of powers is potentially a > good source of checks and balances; some examples: the US > gov't: legislative(bicameral), executive and judicial; > federal, state and local. Domestic and foreign intelligence, > Italy: polizia, carabineri, and military (with several > branches). The benefit of this split is is that decision > making and decision enforcement are separate. In real polities, the separation of powers is an important way to prevent the abuse of power and the rise of dictatorships. So let me stipulate that a similar argument can be made in favor of separation of powers for the governance of online communities. I'm not *completely* persuaded that the differences between meatspace and virtual communities would not undermine the strength of an argument by analogy, but let's suppose the analogy is strong enough. The question then is what powers, in the context of Textop, need to be held separately. Matthew suggests: > Consider an > enforcer who also makes editorial decisions; his decisions > and actions would often result in a far-from transparent > situation, and be difficult to monitor for abuse. We need some more specific examples. After all, it's quite often the case that enforcement and legislation are close together. I mean, suppose it goes like this: there is a *Leviathan* editor/moderator who decides (after much discussion) that things shall be divided up XYZ. He then leaves it up to the rank-and-file *Leviathan* workers to see to it that things are divided up XYZ. He might look in on their work and make some changes and suggestions himself. Someone attempts to undo the editor's work, and that person is dressed-down by other people (not the editor). Which of these various activities is legislation? Which is enforcement? Which activities should be restricted to one class of participant, and which to a different class? On an active, closely-knit online project like we want Textop to be, these will be real questions. So it really doesn't *help* significantly--it doesn't cause any particular ideas to come to my mind--to say that there should be separation of powers. > A couple wild ideas I'd like to throw out: > > 1)Perhaps this would increase collegiality/make it easier to be an > editor: have the software force editors/moderators to reject > some posts that they would have otherwise accepted. So, say 1 > of 20 good contributions has to be rejected. Thus a rejecter > may be more comfortable/accustomed to rejecting, and an > otherwise touchy rejectee will have to expect some rejection. If the purpose of such a policy is simply to make sure that editors do not accept everything that comes down the pike, I think it is safe to say that there's no need for such a policy, because there's *definitely* going to be work that will be rejected or, at least, changed. It all depends on what the *editorial processes* are, i.e., how exactly work is going to be done. Will it basically be a wiki "with playground police"? Or what? Perhaps I'll write about that soon, but don't let me stop if you if you feel inspired. > 2)Since we aren't allowing anonymity or pseudonymity, we will > also need to verify to some extent that alleged Real Names > are indeed real. For example, Thawte web-of-trust > (http://www.thawte.com/secure-email/web-of-trust-wot/index.htm l) could be used for this; claimed university affiliation is generally web-verifiable. I don't think this is worth it. We would lose far too many potential participants by making them do that. We'll just have to confirm a real mailing address, together with some other information that connects them to a real person. But for the most part it will be community expectations that will keep people in line. Of course, we can expect trolls that will try to break the system; but we'll dispatch them with prejudice. Again, it all depends on *editorial processes*--which we haven't established yet. By the way, what you *might* have in mind, when you say that there should be a separation of powers, is that the person(s) who decide to eject someone from the project should not be the same as the editor of group where the ejectee is causing trouble. I think that's probably right, as long as a system can be set up that ejects obvious troublemakers quickly and efficiently. We want to avoid editors abusing their power and setting up their own personal petty bailiwicks, that give credence to only one of many credible views; but, just as much, we want to avoid a Wikipedia-like situation where trolls are able to game the system and drive away valuable contributors. > I'm not confident that "scientific background will be respected by the people." I agree. We've already seen on Wikipedia--and you can see this on many forums online--that many intelligent, prideful thinkers are simply unwilling to defer to *specialists*. Well, Textop isn't a discussion, it's a reference project, and so decisions, and *high-quality* decisions, need to be made. They will be made by specialists, and so those intelligent, prideful thinkers, who (granted) know a thing or two about the subject but aren't specialists, will be expected to defer to the specialists. > Perhaps the project should have a set of basic axioms that contributors are required to agree with, e.g. the scientific method, the illogic of a list of arguments, such as ad hominem (partly mentioned in the manifesto).... Is this a good idea, or is it needlessly divisive? I think some such list could be a good idea, but I'm generally opposed to creating policies needlessly. So what would the list of "axioms" accomplish? If the purpose of a policy can be served by putting the right people in charge, then let's put the right people in charge and forego the policymaking. > Also, if we work hard and create "a meritocracy with enforceable rules" that proves to work very well, we could see the sincerest form of flattery. Indeed. That's my hope. I hope we can be an example to the broader Digital Universe and to the Internet generally. --Larry From larry.sanger at corp.manyone.net Mon May 29 18:51:41 2006 From: larry.sanger at corp.manyone.net (Larry Sanger) Date: Mon, 29 May 2006 18:51:41 -0700 Subject: [Textop] Re: Pilot project proposal: your comments solicited In-Reply-To: <5D2D795A0116244CABFE285440AA48AA01728904@ouebe101.NOE.Nokia.com> Message-ID: <002301c6838b$98be9f00$a0f5fea9@D7WRQ591> All right, so as you can see below, there's one person on record saying that a pilot project for Textop would be good. Anyone else? I think we *could* try the pilot project with essays. It would not be as interesting, in the long run, because the outline that results would not be very long or detailed, but I think that might be a good way forward. The reason I say an outline based on essays would not be very long or detailed is that, according to rules I have in mind (and a DRAFT of which I've articulated here http://www.textop.org/collation_principles.html): "Do not create a new outline node unless (1) there a chunk to place (at least provisionally) under it, or (2) there is a sub-node under the node which is itself created specifically as a place for a chunk." So: "In other words, the outline is not created a priori. It grows exactly where needed." But this doesn't mean that the project using essays wouldn't raise many issues and clarify many things in many people's minds. I'd be in favor of using books of essays, or books that have essay-like chapters (I think of Hume and Russell). --Larry > -----Original Message----- > From: textop-bounce at freelists.org > [mailto:textop-bounce at freelists.org] On Behalf Of > EXT-Mathias.Brendel at nokia.com > Sent: Monday, May 15, 2006 5:34 AM > To: textop at freelists.org > Subject: [textop] Re: Pilot project proposal: your comments solicited > > > Hi all, > > I did not read the description in detail, but I had the same > idea in my subconscious.:) > > I think this is great and we can try out ideas. My English is > not so good, but I can read and understand philosophical > books, and I am even on the edge of publishing some articles > in English. The problem is that my English is really on the > limit of acceptable (for that). Maybe it will be ok for > Textop, since I do not have to produce philosophical essays > here. Moreover, I hope, native speakers can correct me. > > Regarding the target of this project. > > Could we do the pilot project with essays? They are shorter, > anybody can read them if they do not know them. Moreover, > they may be available on-line. So let us first select a > couple of essays from Russell or Pierce or whatever. > > M. > > >-----Original Message----- > >From: textop-bounce at freelists.org > >[mailto:textop-bounce at freelists.org] On Behalf Of ext Larry Sanger > >Sent: Sunday, May 14, 2006 11:39 PM > >To: textop at freelists.org > >Subject: [textop] Pilot project proposal: your comments solicited > > > >All, > > > >After more thought, I have come to the *tentative* view that > >we should use a wiki to do some pilot projects. Please see my > >proposal here: > > > >http://www.textop.org/pilot_project.html > > > >*Every aspect* of this pilot project proposal is up for > >discussion. Feel free to offer feedback on the list, or to me > >personally (if you feel more comfortable that way). > > > >--Larry > > > >== > >textop - a Textop (http://www.textop.org) mailing list. > >To post, send a mail to textop at freelists.org, or just reply to > >this post. > >To unsubscribe, send a mail to textop-request at freelists.org > >with 'unsubscribe' in the header. > > > == > textop - a Textop (http://www.textop.org) mailing list. > To post, send a mail to textop at freelists.org, or just reply > to this post. To unsubscribe, send a mail to > textop-request at freelists.org with 'unsubscribe' in the header. > From larry.sanger at corp.manyone.net Mon May 29 23:19:39 2006 From: larry.sanger at corp.manyone.net (Larry Sanger) Date: Mon, 29 May 2006 23:19:39 -0700 Subject: [Textop] Signing up In-Reply-To: <447701F3.5010106@starpower.net> Message-ID: <003201c683b1$08851fb0$a0f5fea9@D7WRQ591> Hi Mark, You don't need to sign up again. Sorry for the delayed response. --Larry > -----Original Message----- > From: textop-bounces at lists.dufoundation.org > [mailto:textop-bounces at lists.dufoundation.org] On Behalf Of M Flood > Sent: Friday, May 26, 2006 6:26 AM > To: textop at lists.dufoundation.org > Subject: [Textop] Signing up > > > Greetings all: > > I have a simple administrative question that may be of > interest to others. (I hope I'm not the only one who can't > figure this out!) > > I signed up several months ago to learn more about the > stewardship program. Now that there's been an official call > for participation, I looked at the new recruitment page > (www.dufoundation.org/stewards.php), > which has a sign-up form that appears slightly different from > what I used before. > > Should I sign up again, or am I already in the database? > > Thanks -- Mark Flood _______________________________________________ > Textop mailing list > Textop at lists.dufoundation.org > http://lists.dufoundation.org/mailman/listinfo/textop > From math at freemail.hu Tue May 30 05:34:27 2006 From: math at freemail.hu (Brendel Matyas) Date: Tue, 30 May 2006 14:34:27 +0200 (CEST) Subject: [Textop] Pilot project proposal: your comments solicited In-Reply-To: <002301c6838b$98be9f00$a0f5fea9@D7WRQ591> Message-ID: Hi all, > But this doesn't mean that the project using essays wouldn't raise many > issues and clarify many things in many people's minds. I'd be in favor of > using books of essays, or books that have essay-like chapters (I think of > Hume and Russell). Lets have this as a compromissum. Is this OK for everyone? Bertrand Russell, Problems of Philosophy I am afraid "everyone" means Larry and me.:) Matthias _______________________________________________________________________________ Szem?lyi k?lcs?nt interneten kereszt?l, hitelb?r?lati d?j n?lk?l, ak?r 72 h?napos futamid?re! http://www.klikkbank.hu/lakossagi/termekek/szemelyi_kolcson/index.html From davcormier at upei.ca Tue May 30 11:08:53 2006 From: davcormier at upei.ca (dave cormier) Date: Tue, 30 May 2006 15:08:53 -0300 Subject: [Textop] Pilot project proposal: your comments solicited In-Reply-To: References: Message-ID: <447C8A35.3050204@upei.ca> Sounds like a fine place to start... mostly because i've read it... which makes it easier. Brendel Matyas wrote: >Hi all, > > > >>But this doesn't mean that the project using essays wouldn't raise >> >> >many > > >>issues and clarify many things in many people's minds. I'd be in favor >> >> >of > > >>using books of essays, or books that have essay-like chapters (I >> >> >think of > > >>Hume and Russell). >> >> > > >Lets have this as a compromissum. > >Is this OK for everyone? > >Bertrand Russell, Problems of Philosophy > >I am afraid "everyone" means Larry and me.:) > >Matthias > > >_______________________________________________________________________________ >Szem?lyi k?lcs?nt interneten kereszt?l, hitelb?r?lati d?j n?lk?l, ak?r 72 h?napos futamid?re! >http://www.klikkbank.hu/lakossagi/termekek/szemelyi_kolcson/index.html > > >_______________________________________________ >Textop mailing list >Textop at lists.dufoundation.org >http://lists.dufoundation.org/mailman/listinfo/textop > > > -- dave cormier Project Manager, Digital Library Tech Coordinator, CMTC phone: 902-566-6007 email: davcormier at upei.ca skype: coarsesalt From larry.sanger at corp.manyone.net Tue May 30 11:15:31 2006 From: larry.sanger at corp.manyone.net (Larry Sanger) Date: Tue, 30 May 2006 11:15:31 -0700 Subject: [Textop] Pilot project proposal: your comments solicited In-Reply-To: <447C8A35.3050204@upei.ca> Message-ID: <001f01c68415$099be870$a0f5fea9@D7WRQ591> All, This sounds fine with me. I would caution that we'll want to have people who actually have substantial background in philosophy--many classes, at the very least--and who have read the book, working on collating it. Also, we'll want to find a Russell expert to serve as editor/overseer. What do you think? --Larry > -----Original Message----- > From: textop-bounces at lists.dufoundation.org > [mailto:textop-bounces at lists.dufoundation.org] On Behalf Of > dave cormier > Sent: Tuesday, May 30, 2006 11:09 AM > To: General Discussion List for the Text Outline Project > Subject: Re: [Textop] Pilot project proposal: your comments solicited > > > Sounds like a fine place to start... mostly because i've read it... > which makes it easier. > > Brendel Matyas wrote: > > >Hi all, > > > > > > > >>But this doesn't mean that the project using essays wouldn't raise > >> > >> > >many > > > > > >>issues and clarify many things in many people's minds. I'd be in > >>favor > >> > >> > >of > > > > > >>using books of essays, or books that have essay-like chapters (I > >> > >> > >think of > > > > > >>Hume and Russell). > >> > >> > > > > > >Lets have this as a compromissum. > > > >Is this OK for everyone? > > > >Bertrand Russell, Problems of Philosophy > > > >I am afraid "everyone" means Larry and me.:) > > > >Matthias > > > > > >_____________________________________________________________ > __________ > >________ > >Szem?lyi k?lcs?nt interneten kereszt?l, hitelb?r?lati d?j > n?lk?l, ak?r 72 h?napos futamid?re! > >http://www.klikkbank.hu/lakossagi/termekek/szemelyi_kolcson/i > ndex.html > > > > > >_______________________________________________ > >Textop mailing list > >Textop at lists.dufoundation.org > >http://lists.dufoundation.org/mailman/listinfo/textop > > > > > > > > > -- > dave cormier > Project Manager, Digital Library > Tech Coordinator, CMTC > > phone: 902-566-6007 > email: davcormier at upei.ca > skype: coarsesalt > > _______________________________________________ > Textop mailing list > Textop at lists.dufoundation.org > http://lists.dufoundation.org/mailman/listinfo/textop > From davcormier at upei.ca Tue May 30 11:32:19 2006 From: davcormier at upei.ca (dave cormier) Date: Tue, 30 May 2006 15:32:19 -0300 Subject: [Textop] Pilot project proposal: your comments solicited In-Reply-To: <001f01c68415$099be870$a0f5fea9@D7WRQ591> References: <001f01c68415$099be870$a0f5fea9@D7WRQ591> Message-ID: <447C8FB3.8090405@upei.ca> ok then a few quick thoughts... That becomes the first question I suppose. Do you just expect people to pony up and say... yeah, well, I'm not much of a philosopher. I've got an MEd, and my BA was in philosophy. I also tend a little more toward Derrida than Russell. Now, I've read a bunch of Russell, and happen to own two copies of the book at issue. Do you have a place for CVs yet? It might also be helpful to have a very clear description, at the start of each... module... node... as to what qualifies as an expert for that section. And a system of roles as well... While my resume might not qualify me for content creation, my other experience might qualify me in another role. -- dave cormier Project Manager, Digital Library Tech Coordinator, CMTC phone: 902-566-6007 email: davcormier at upei.ca skype: coarsesalt From larry.sanger at corp.manyone.net Tue May 30 11:51:25 2006 From: larry.sanger at corp.manyone.net (Larry Sanger) Date: Tue, 30 May 2006 11:51:25 -0700 Subject: [Textop] Pilot project proposal: your comments solicited In-Reply-To: <447C8FB3.8090405@upei.ca> Message-ID: <013d01c6841a$0d49bbf0$a0f5fea9@D7WRQ591> We do need to speak frankly about this issue. I think our first concerns will be (1) to get a quorum of people actually working on the Collation Project, and (2) to expose a large number of people to the concepts behind the project. Both of these concerns require that we "set the bar low" so to speak. In fact, I have had it in mind that people working on a text should have training to the bachelor's degree level in the subject that traditionally studies the text. For purposes of the pilot project, we might even say the bar is set lower, but let's see who we have on board who's interested in *actually doing work*. :-) Anyway, these wouldn't be billed as "experts" but as regular old collaborators. Bear in mind that, in the future, all decisions about minimum qualifications to do this or that must be made in the service of the reference that we are building--not in the service of our egos. We *might* discover that, to be minimally usable, people really need training beyond the bachelor's level. But I think we should start there. More restrictive decisions will be made by a future executive committee (which doesn't exist yet). Quite frankly, if we discover (much to our surprise) that boatloads of the world's greatest experts on this and that text are interested in contributing to the project, then the bar might be raised very high indeed. That's the nature of a genuine meritocracy. But I don't expect this to happen anytime soon. Still...who knows? As to a place for CVs--well, I haven't set up anything for Textop yet. Don't worry about that yet--basically, if you're interested, make sure you're on one of the mailing lists. --Larry > -----Original Message----- > From: textop-bounces at lists.dufoundation.org > [mailto:textop-bounces at lists.dufoundation.org] On Behalf Of > dave cormier > Sent: Tuesday, May 30, 2006 11:32 AM > To: General Discussion List for the Text Outline Project > Subject: Re: [Textop] Pilot project proposal: your comments solicited > > > ok then a few quick thoughts... > > That becomes the first question I suppose. Do you just expect > people to > pony up and say... yeah, well, I'm not much of a philosopher. > I've got > an MEd, and my BA was in philosophy. I also tend a little more toward > Derrida than Russell. Now, I've read a bunch of Russell, and > happen to > own two copies of the book at issue. > > Do you have a place for CVs yet? It might also be helpful to > have a very > clear description, at the start of each... module... node... > as to what > qualifies as an expert for that section. > > And a system of roles as well... While my resume might not qualify me > for content creation, my other experience might qualify me in > another role. > > -- > dave cormier > Project Manager, Digital Library > Tech Coordinator, CMTC > > phone: 902-566-6007 > email: davcormier at upei.ca > skype: coarsesalt > > _______________________________________________ > Textop mailing list > Textop at lists.dufoundation.org > http://lists.dufoundation.org/mailman/listinfo/textop > From cllawrence at eosceres.com Tue May 30 10:18:54 2006 From: cllawrence at eosceres.com (chris) Date: Tue, 30 May 2006 12:18:54 -0500 Subject: [Textop] Pilot project proposal: your comments solicited In-Reply-To: <447C8FB3.8090405@upei.ca> References: <001f01c68415$099be870$a0f5fea9@D7WRQ591> <447C8FB3.8090405@upei.ca> Message-ID: <447C7E7E.4050401@eosceres.com> An HTML attachment was scrubbed... URL: http://lists.dufoundation.org/archive/textop/attachments/20060530/b02d9070/attachment-0001.htm From larry.sanger at corp.manyone.net Tue May 30 12:45:06 2006 From: larry.sanger at corp.manyone.net (Larry Sanger) Date: Tue, 30 May 2006 12:45:06 -0700 Subject: [Textop] Pilot project proposal: your comments solicited In-Reply-To: <447C7E7E.4050401@eosceres.com> Message-ID: <001101c68421$8d7acba0$a0f5fea9@D7WRQ591> Welcome, Chris! If you're looking for something cross-disciplinary in English, in essay form, that is broadly philosophical, we would do well to start with Bacon's Essays. http://ourworld.compuserve.com/homepages/mike_donnelly/bacon.htm (not necessarily the source we'll want to use, but an example) The trouble with Derrida is that he's still copyrighted. He's also going to be very difficult to "collate" due to the difficulty of deciphering exactly what he's after. Moreover, he's the sort of philosopher who only makes sense in context of and in reaction to other thinkers (he's Mr. Deconstruction, after all)--and so it will be very useful to have an outline structure in advance of trying to tackle Derrida. The other problem with Derrida is that it's in French, and one thing I'm sure the Collation Project will want to do is to set up a procedure whereby texts are collated in their original languages--so when we do tackle Derrida it will be a group of French-speaking scholars who do so. See: http://www.textop.org/internationalization.html http://www.textop.org/reqs_v1.html --Larry -----Original Message----- From: textop-bounces at lists.dufoundation.org [mailto:textop-bounces at lists.dufoundation.org] On Behalf Of chris Sent: Tuesday, May 30, 2006 10:19 AM To: General Discussion List for the Text Outline Project Subject: Re: [Textop] Pilot project proposal: your comments solicited My first post. I think I must be the only woman on here also! I would agree with Dave because I think writings from someone like Derrida might cross disciplines more. I'm not sure I would choose Derrida, but his work is a good start. My expertise is in Educational Philosophy (PhD in Educational Psychology) - Derrida is often cited in research in Educational Philosophy. I also have expertise in the nature and philosophy of science/science education (many excellent authors in this field also that might cross disciplines) and qualitative research. I know Bertrand Russell writings are intriguing and cross disciplines, but I think other authors works would be a better place to start. Just some thoughts (and congrats on the new baby Larry!) Chris Lawrence dave cormier wrote: ok then a few quick thoughts... That becomes the first question I suppose. Do you just expect people to pony up and say... yeah, well, I'm not much of a philosopher. I've got an MEd, and my BA was in philosophy. I also tend a little more toward Derrida than Russell. Now, I've read a bunch of Russell, and happen to own two copies of the book at issue. Do you have a place for CVs yet? It might also be helpful to have a very clear description, at the start of each... module... node... as to what qualifies as an expert for that section. And a system of roles as well... While my resume might not qualify me for content creation, my other experience might qualify me in another role. -- ~~~~~~<.*.>~~~~~~ my art Every journey has a secret destination of which the traveler is unaware. ~Martin Buber -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.dufoundation.org/archive/textop/attachments/20060530/5477801e/attachment.htm From cllawrence at eosceres.com Tue May 30 10:54:40 2006 From: cllawrence at eosceres.com (chris) Date: Tue, 30 May 2006 12:54:40 -0500 Subject: [Textop] Pilot project proposal: your comments solicited In-Reply-To: <013d01c6841a$0d49bbf0$a0f5fea9@D7WRQ591> References: <013d01c6841a$0d49bbf0$a0f5fea9@D7WRQ591> Message-ID: <447C86E0.5020906@eosceres.com> An HTML attachment was scrubbed... URL: http://lists.dufoundation.org/archive/textop/attachments/20060530/4d063748/attachment.htm From ghburrows at comcast.net Tue May 30 12:59:15 2006 From: ghburrows at comcast.net (Howard Burrows) Date: Tue, 30 May 2006 15:59:15 -0400 Subject: [Textop] Textop wiki set up In-Reply-To: <000001c68200$eae22d60$a0f5fea9@D7WRQ591> Message-ID: <004901c68423$87a272d0$6400a8c0@Quacky> Ruth and I and one other anonymous party are participating on the textop wiki. The wiki allows room for CV materials and statements of interest, as everyone gets their own public registration page. So far most of the discussion is focusing on how to use the wiki and can be found by clicking the "discussion" tab attached to Larry's Main Page. Howard Burrows -----Original Message----- From: textop-bounces at lists.dufoundation.org [mailto:textop-bounces at lists.dufoundation.org] On Behalf Of Larry Sanger Sent: May 27, 2006 Saturday 10:46 PM To: textop at lists.dufoundation.org Subject: [Textop] Textop wiki set up http://www.textop.org/wiki All, I've set up MediaWiki in the Textop space; see above. I've put up the pilot project proposal on the main page, ready for you to edit (or comment on, here: http://www.textop.org/wiki/index.php?title=Talk:Main_Page&action=edit). I'll be getting caught up on textop mailing list mail soon...in the meantime, the question foremost in my mind now is whether we ought to try out a pilot project soon, as outlined above, or whether we should (for some reason) postpone the pilot project. --Larry _______________________________________________ Textop mailing list Textop at lists.dufoundation.org http://lists.dufoundation.org/mailman/listinfo/textop From larry.sanger at corp.manyone.net Tue May 30 16:49:58 2006 From: larry.sanger at corp.manyone.net (Larry Sanger) Date: Tue, 30 May 2006 16:49:58 -0700 Subject: [Textop] FW: still not fine-grained and structured enough to bescalable In-Reply-To: <200605191521.k4JFLRap011429@pc070372.sci.griffith.edu.au> Message-ID: <001801c68443$c2ad7260$7e02020a@D6WD1391> First, I would like to thank Philippe for his ongoing expert thinking and work on the relationship between his work and Textop. I am a great supporter of work like his--formal ontology--and maybe someday I will explore that field myself. Also, I think it is very important that we (many of us) understand what relation properly should exist between the Textop projects and formal ontology. I don't think I personally have really "grokked" that yet myself. Philippe Martin wrote: > So I think these two implementations/approaches are complementary and that they should point to each other, to permit people who want to find and update more (vs. less) fine grained info units than paragraphs and topics (vs. sentences and concepts) to find an adequate site. These people may even participate to both (but it is clear that most persons will prefer Larry's approach). I agree with all this (and am happy Philippe found my article worthy of mention). > There are also a lot of possible cross-overs between the two approaches: I allow paragraphs, and the relations I propose could be re-used in the TextOp or Debate projects. So I consider my work as part of the TextOp Pilot project and hence would like you to consider the next 2 documents: http://www.webkb.org/kb/classif/sdLeviathan.html http://www.webkb.org/kb/classif/sdGodExistence.html The first represents/indexes statements and concepts from the first five chapters of the Leviathan, This is very interesting, and we ought to link to your work somehow--probably on the wiki. (More on that later as I respond to Howard Burrows.) As to your representation of "statements and concepts from the first five chapters of the Leviathan," I want to make an observation. *Completely* mapping out the logical structure of all statements in the first five chapters would require a great deal more work than you have done here, as I'm sure you (Philippe) are well aware. Doing that *would* be very interesting, but it would also be very difficult. Furthermore, I think the Collation Project would be an excellent preparative to your project. By doing more coarse-grained chunking of texts (paragraphs rather than sentences or parts of sentences), we will still achieve remarkable detail and specificity in our outlining, but we will avoid the inevitable difficulties in coming to an agreement about, and teaching of, the principles of constructing a formal representation of a text. It is precisely the relative "sloppiness" of the Collation Project that makes it more tractable. Furthermore, the end result--relatively self-contained paragraph-length chunks of text put side-by-side with other chunks on the same topic--will be easier to understand, use, and learn from than formal representations of content. > and the second represents and complements the argumentation tree used > in John Locke's Chapter 9 of his essay "Of our Threefold Knowledge of Existence". The interface to access and browse those representations from the lexical ontology of WebKB-2 and from related representations will be available in July (it will basically be the current interface of WebKB-2 but with more presentation options). This is also interesting--reminds me of many handouts I used to give my students, summarizing arguments of philosophers. So, though I definitely see the value of formal representations of arguments (having done this quite a bit myself), I am not sure that that will be the most useful way to present arguments in the Debate Guide Project--but let me emphasize that I am *very* open to debate on this subject! As to the rest: I will make an effort to read http://www.webkb.org/doc/papers/iccs05/iccs05.html soon, as it is very closely related to our task. > My problem is that I fail to imagine how your mono-source example > generalise into one structure that organises the topics and arguments > of thousands of documents, without using an ontology to solve the > problem of information organisation and retrieval. > I certainly agree that there will be a lot of redundancies in the > sources that people will exploit and hence they can for example choose not to represent arguments that are equivalent to arguments that have previously been represented. However, for each idea/statement there will still be a *lot* of arguments, sub-arguments, counter-arguments on the sub-arguments, etc. How will they be organised (and then compared) if specialization and argumentation relations are not used? The answer is that philosophers, legal scholars, and many others are already very well-trained in making distinctions and at least rough classifications of different arguments, theories, principles, etc. In doing this work, they do of course represent (not necessarily formally) arguments, sub-arguments, counter-arguments, etc. as such. I think, again, of the many handouts I've given students, providing just such a map of a conceptual territory. In fact, the *more* data there is, the better the map will be, I think; having just one source, like the Leviathan, tends to skew things. > It may be that you do not want the sub-arguments to be organised ("the > items I propose to classify are these sorts of 'text chunks'") but then (i) information retrieval and comparison may well be an issue, (ii) it would clearly be "of more interest" to organize them (if people are willing to do that) even though your approach is already very interesting. Think of it rather like this: a text chunk will be filed according to *what its author appears to be doing with it*. So if in a chunk--an approximately paragraph-length part of the text--the author is discussing just a special portion of an argument, then that chunk is filed in a subheading. If the chunk is a summary of the whole argument, then it is filed under the parent heading. How deeply a piece of text is filed depends on how specialized the focus of the text is. The exception to this, I imagine, is when scholars recognize as a seminal text about some relatively specialized discussion that the author doesn't focus on--a comment made in passing. Then collators might prise out the relevant text in order to place the later commentaries on it in context. > Furthermore, regarding "topics" now, how the tens or hundreds of thousands of topics that the text chunks are about will be organized (and hence retrieved) if specialization and other relations are not explicitely used between these concepts/topics? Why wouldn't specialization and other relations be explicitly used between topics? I'm not sure why you assume they wouldn't be. Explicit use of rules about node relationships is possible without requiring that the results being formalizable. There can be "slop," if formalization isn't the aim. > A Yahoo/DMOZ like hierarchy of topics would be hopeless (if we do not agree on that, please tell me but I believe we do). I agree entirely. > Keyword-based retrieval would be hopeless too. You are talking about relations in your answer ("...designate what the parent-child relations shall mean... more definite relations ... in relation to a cluster of other related concepts ... what relation a specific point about realism ...) but not what kinds of relations nor if they would always be explicited. Yes, I think it would be a good idea, anyway, to make relations explicit. But not at the expense of a coherent, nuanced outline. I mean, we don't want to try to "shoehorn" (force) ourselves using a set of rules...rather, the rules should come out of our experience with actually constructing a good outline. We can start with some definite ideas, but they must be refined by experience. This isn't the sort of thing that can be done a priori. > > And how we can expect people constructing the outline just to buy into > > the ontology wholesale? > Please tell me/us what kinds of relations you envisage between the text chunks and I think I will undertand your approach. See: http://www.textop.org/collation_principles.html Particularly, "Abstract outline division rules." I have another idea to add to this, but it is just a rather lame theory that seems ad hoc to me. It involves selecting a main topic, such as "human being" or "expression" or "law," and then setting out (1) definitions (and debates about definitions, etc.), (2) essential concepts (i.e., concepts taken from the definition that haven't appeared yet in the outline, could include types or parts), (3) functions (i.e., what the entity of the topic is "for," what its essential activity is), and (4) activities (i.e., the various ways the entity of the topic acts or can be manifested). Any concept from any part of this four-part structure can itself be made into another "main topic," so that the structure then iterates. This is actually just a lame attempt to understand what I was doing with my outline of the Leviathan. I tried to rewrite my outline so that it fits in this structure, but never finished (I'm not sure it can be done). I'm sure we'll come up with something much better as we do more work and explore the possibilities. > Again, if the relations are explicit and have a known meaning, this is an ontology (strictly speaking, even a list of terms is an ontology but I am raising the bar a little bit). If your own experience tells you that explicit relations are not needed for your project, I do not think there is a way I can convince you. For what you have in mind, explicit relations may actually not be needed. I agree that explicit relations are good, but what they should be will depend on the data we have to work with and the best ideas we all can produce for representing it in outline form. > > How would *the collation of texts* proceed? > > What's the overall procedure? > My answer to this kind of question is generally: choose any methodology you wish or none at all, just make sure you respect (or make the users respect) the semantics of the relations they use, this will permit an incremental and correct building of the knowledge base (I also assume cooperation protocols such as those I propose if there are many knowledge providers). This certainly worked for me each time I represented knowledge and, although I acknowledge that knowledge acquisition methodologies are interesting, I have not seen any ontology engineering methodology that really was doing anything else that providing a set of top-level concepts and relations, and giving advices on how to use them. > Surprisingly (for me), so far the above answer rarely satisfied the persons who asked that kind of question. I doubt this will be the case this time but I still hope to. However, I can provide the merged set of the top-level concepts and relations advocated by many methodologies, including argumentation methodologies. For organising statements, I found the following relations useful: specialisation, correction, corrective_specialization, more_concise_reformulation, more_precise_reformulation, objection, argument, example, opposition and consistent_with. I have an idea. Why don't you use (if you can) the wiki to give your treatment to the texts that we choose, and we can compare results? I'd like make just one other point about your approach, the very fine-grained approach. If the logical structure of particular arguments (and other types of text) are to be laid out as you have it, then how are many texts supposed to be *collated*? It's easy to see how the screenshot (http://www.textop.org/screenshot.html) or the sample (http://www.textop.org/outline_help.html) could be augmented with many texts from many sources. But it's not easy to see how your examples http://www.webkb.org/kb/classif/sdLeviathan.html http://www.webkb.org/kb/classif/sdGodExistence.html could easily be expanded with texts from multiple sources. The scalability of the Collation Project is pretty obvious in my mind, but the scalability of yours, I just don't really get. --Larry From phmartin at phmartin.info Tue May 30 22:37:37 2006 From: phmartin at phmartin.info (Philippe MARTIN) Date: Wed, 31 May 2006 15:37:37 +1000 Subject: [Textop] FW: still not fine-grained and structured enough to be scalable In-Reply-To: Your message of "Tue, 30 May 2006 16:49:58 MST" Message-ID: <200605310537.k4V5bgqv002747@pc070372.sci.griffith.edu.au> Larry, > *Completely* mapping out the logical structure of all statements in the > first five chapters would require a great deal more work than you have > done here, as I'm sure you (Philippe) are well aware. Oh yes, as noted in my document (in which I give my selection criteria). As also noted, there are a lot of statements in the source text that are not "logicaly"/"argumentatively" related. > ... a formal representation of a text ... Again, I'd like to point out that my examples were very far from being "formal" (since in those examples the connected statements are informal; only the relations between them are explicit and catalogued in my ontology). For lack of a better term, my approach is "semi-formal". (Of course, my tool also permits courageous users to enter "formal" statements). > In fact, the *more* data there is, the better the map will be, I think I do not think I have encountered a counter-example to that rule, whatever activity/approach and data I used (e.g., the bigger a software, the more structured it tends to be and also "has to" be). > ... reminds me of many handouts I used to give my students, summarizing > arguments of philosophers. I'd be very interested if you could email me some of these documents or, even better, an URL where to find them. > ... and then setting out (1) definitions (and debates > about definitions, etc.), (2) essential concepts (i.e., concepts taken > from the definition that haven't appeared yet in the outline, could > include types or parts), (3) functions (i.e., what the entity of the > topic is "for," what its essential activity is), and (4) activities I similarly have to group my relations from an object (say, one activity) into clusters (for which one could use the headings: "who", "when", "for what", "what", "with what", ...) when there are a lot of relations. (And so on for the sub-activities). This can be used both for presentation or methodological purposes. > It's easy to see how the > screenshot (http://www.textop.org/screenshot.html) or the sample > (http://www.textop.org/outline_help.html) could be augmented with many > texts from many sources. But it's not easy to see how your examples > http://www.webkb.org/kb/classif/sdLeviathan.html > http://www.webkb.org/kb/classif/sdGodExistence.html > could easily be expanded with texts from multiple sources. The > scalability of the Collation Project is pretty obvious in my mind, but > the scalability of yours, I just don't really get. This is (amusingly and unfortunately) a rephrasing of sentences from my original message but with an inversion of the roles. "Unfortunately" because each of us finds his own vision so obvious that we do not know why the other does not find it scalable (and "the most scalable") and hence cannot find better explanations. But again, this may be only a matter of a difference about goals/purposes. In any case, we have tried (and this led to several precisions and agreements), and now only further examples may bring more light. > I have an idea. Why don't you use (if you can) the wiki to give your > treatment to the texts that we choose, and we can compare results? If I have time, yes I'd like to. (No doubt regarding the possibility to tackle any subject with my approach). To test some of the scalability issues, there should be several texts about the same kinds of ideas. "compare results" seems to indicate that the semi-formal representations will be separated from the other (informal but still structured) representations. If this is not the case, please precise. Philippe From larry.sanger at corp.manyone.net Wed May 31 17:58:58 2006 From: larry.sanger at corp.manyone.net (Larry Sanger) Date: Wed, 31 May 2006 17:58:58 -0700 Subject: [Textop] Role-play-game In-Reply-To: <5D2D795A0116244CABFE285440AA48AA0179D577@ouebe101.NOE.Nokia.com> Message-ID: <003901c68516$908be5e0$a0f5fea9@D7WRQ591> Still catching up with old mail here. Now from Mathias Brendel: > I noticed that I had a point unfinished. This is because I > noticed that I will explain it later. But I did not delete > the unfinished part. > > So > > "For me it is reasonable at present that there shall be at > least 2 roles beside the collaborator. One is the editor. He has > power to reject a piece of work to be published. But he has no" > > The explanation fotr this was this: > > "An editor is somebody, who has control over publishing > something. But he shall not have power over people. The > administrator can ban somebody (details to be worked out), but > shall not have power over editorial questions." > > So these are 2 important roles in my opinion. There is sure a > role of a general collaborator. This is clearer: so, that the "separation of powers" discussed earlier comes down to (at least) the claim that editors do not have the power to ban people. > Further, there could be a role of arbitrator. We did not have > this in the Hungarian Wikipedia, and I think, this is > important. If two parties are quareling, it is important to > have a third point of view. The arbitrator shall be > independent. Here again, separation of roles is important. > There is no use of an arbitrator, if he is also an administrator. While perhaps we could use *a* role of arbitrator or ombudsman, I think that Wikipedia's system is broken here. Frequently, people are asked to go into "arbitration" with people who are simply unreasonable--which is a waste of time for productive users who have been made the target of ideologues and trolls. I think we need a more efficient and effective way to identify and remove troublemakers. I'm particularly opposed to a policy where arbitration between individuals is required as a condition of further action. > Checkuser: this was a role in Wikipedia, which I do not think > is good. Actually the problem is not with the role, but with > the question of anonymity. I think, we shall not allow anonym > collaborators. Moreover, I think it would be desired that > every collaborator takes part with his real name. We shall > know him/her. That means that we shall know his/her > background. I have no detailed idea about this. I can imagine > many solution. Shall we require some kind of a CV? In any > case, a collaborator with a good background shall get high > respect in the project. What information we should require from participants, as a condition of their participation, is a big and difficult question. Generally, we have to weigh three important considerations: (1) the ease with which a person can get involved (which is very important when we're dealing with volunteers), (2) having some guarantee of the minimum qualifications of participants, and (3) our ability to rid the project of troublemakers. I think that simply requiring people to use their own real names will go a long way to solving a lot of problems--this actually is the Digital Universe Foundation view. I think that we will probably want, in addition, *some* piece of information that we can use to do "due diligence" in determining that a person is who he says he is, although I also think we should be flexible here. I think that, basically, if you design your policies around fear of interference, you'll create a walled garden that goes nowhere. In practical terms, we'll want a very simple "application" system in which editors can view basic information a person volunteers about him- or herself, and then decide whether the person should be able to work on a given text. There's much more to be said here--these are only provisional first thoughts. --Larry