From nikodemus at random-state.net Fri Nov 10 12:44:20 2006 From: nikodemus at random-state.net (Nikodemus Siivola) Date: Fri, 10 Nov 2006 14:44:20 +0200 Subject: [cdr-devel] Mailing lists, document licences Message-ID: <87lkmjii8r.fsf@logxor.random-state.net> First of all, congratulations on CDR getting some traction! Secondly, what's the charter of the cdr-devel list? Is it only for "talking about the CDR process", or are existing CDRs a proper topic too? If not, I propose a separate list cdr-talk, for which existing CDRs are on topic, and where the subject line should carry a [CDR N] header to indicate which CDR is addressed. (Leaving the cdr-devel for meta-discussions on the process and whatnot.) Also, can we get the lists (including the archives) on Gmane, please? The subscription form is here: http://gmane.org/subscribe.php Finally, have you considered providing a liberal boilerplate licence (eg. CC attribution only, or 1-clause MIT) that CDR authors could use should they want to? I am cogitant of the fact that such a licence may not be appropriate for all CDRs, but for those for which it is, it would be nice to have a single default choise to cut down on possible future ambiguiety. Cheers, -- Nikodemus Schemer: "Buddha is small, clean, and serious." Lispnik: "Buddha is big, has hairy armpits, and laughs." From pc at p-cos.net Fri Nov 10 14:07:35 2006 From: pc at p-cos.net (Pascal Costanza) Date: Fri, 10 Nov 2006 15:07:35 +0100 Subject: [cdr-devel] Mailing lists, document licences In-Reply-To: <87lkmjii8r.fsf@logxor.random-state.net> References: <87lkmjii8r.fsf@logxor.random-state.net> Message-ID: <9DFA04C9-1532-424B-A86A-E39072C8C66B@p-cos.net> Hi Nikodemus, Thanks a lot for the congratulations ;) and also for the suggestions. Before I comment on specific details of your suggestions, here is first a general disclaimer: The idea behind CDR is to be as light- weight as possible. We have tried to ensure that managing the CDR process is very straightforward and doesn't involve too much overhead. It's our impression that such processes can easily create too much overhead and requires work that doesn't really help improve the contents of the CDRs themselves. That's why some ideas that we have discussed ourselves didn't get into the CDR process because of potential hidden harmful costs. So for example, we specifically didn't add any way to "approve" documents, or prescribe a format, etc., because this could put us in the position of having to judge something that we actually cannot judge, which would block the whole process. So much for the background of the CDR process. So now to your actual suggestions. (The following are my personal opinions and don't necessarily reflect those of the other CDR editors.) On 10 Nov 2006, at 13:44, Nikodemus Siivola wrote: > First of all, congratulations on CDR getting some traction! > > Secondly, what's the charter of the cdr-devel list? Is it only for > "talking about the CDR process", or are existing CDRs a proper topic > too? > > If not, I propose a separate list cdr-talk, for which existing CDRs > are on topic, and where the subject line should carry a [CDR N] header > to indicate which CDR is addressed. (Leaving the cdr-devel for > meta-discussions on the process and whatnot.) So far, the mailing list is only for discussing the CDR process itself, not for discussing specific CDRs. There are different ways to produce and discuss documents, not only through mailing lists. We don't want to require any specific way how people come up with CDRs, we don't want to require their public discussion in mailing lists, and we don't to give the impression that this is the case. We also don't want to give the impression that CDRs are "approved" by the participants in mailing lists. There are several options to create mailing lists, install Wikis, or whatever, outside of the CDR process, and we would like to encourage authors to use whatever they think is most appropriate. If authors/ submitters haven't done so, you are still able to contact them directly to send them your feedback on specific CDRs. Maybe there is a middle ground here that I simply don't see at the moment. But I think that our encouragement in the CDR manual to have documents publicly discussed before they are submitted is sufficient. > Also, can we get the lists (including the archives) on Gmane, > please? The subscription form is here: > > http://gmane.org/subscribe.php Do you think that cdr-level is appropriate for gmane when it is only about the CDR process? (That's not a rhetorical question. ;) > Finally, have you considered providing a liberal boilerplate licence > (eg. CC attribution only, or 1-clause MIT) that CDR authors could use > should they want to? I am cogitant of the fact that such a licence may > not be appropriate for all CDRs, but for those for which it is, it > would be nice to have a single default choise to cut down on possible > future ambiguiety. ...but there are enough resources on the net where you can find such licenses. Again, the CDR process intentionally doesn't involve any quality control, and we also don't want to favor open source approaches over, say, commercial software vendors. Any specific recommendation could give the wrong impression here. Sorry that this email sounds as if we simply want to reject all your suggestions, but that's certainly not the case. Cheers, Pascal -- Pascal Costanza, mailto:pc at p-cos.net, http://p-cos.net Vrije Universiteit Brussel, Programming Technology Lab Pleinlaan 2, B-1050 Brussel, Belgium From nikodemus at random-state.net Fri Nov 10 18:48:39 2006 From: nikodemus at random-state.net (Nikodemus Siivola) Date: Fri, 10 Nov 2006 20:48:39 +0200 Subject: [cdr-devel] Mailing lists, document licences References: <87lkmjii8r.fsf@logxor.random-state.net> <9DFA04C9-1532-424B-A86A-E39072C8C66B@p-cos.net> Message-ID: <87k623nnnc.fsf@logxor.random-state.net> Pascal Costanza writes: > Before I comment on specific details of your suggestions, here is > first a general disclaimer: The idea behind CDR is to be as light- > weight as possible. --snip-- > So for example, we specifically didn't add any way to "approve" > documents, or prescribe a format, etc., because this could put us in > the position of having to judge something that we actually cannot > judge, which would block the whole process. ...which is indeed one of the reasons why I am so enthusiastic about its possibilities. ;-) > So far, the mailing list is only for discussing the CDR process > itself, not for discussing specific CDRs. There are different ways > to produce and discuss documents, not only through mailing lists. We > don't want to require any specific way how people come up with CDRs, > we don't want to require their public discussion in mailing lists, > and we don't to give the impression that this is the case. We also > don't want to give the impression that CDRs are "approved" by the > participants in mailing lists. > > There are several options to create mailing lists, install Wikis, or > whatever, outside of the CDR process, and we would like to encourage > authors to use whatever they think is most appropriate. If authors/ > submitters haven't done so, you are still able to contact them > directly to send them your feedback on specific CDRs. > > Maybe there is a middle ground here that I simply don't see at the > moment. But I think that our encouragement in the CDR manual to have > documents publicly discussed before they are submitted is sufficient. I see your point, but I do think there is a middle ground to be found: while document authors are free to choose any venue of their liking for discussions about their document, it would be handy if a "sensible default" existed. Quoting from the CDR website: "You are encouraged, but not required to: * Discuss your document publicly before you submit it as a CDR document, for example in mailing lists, newsgroups, or other public forums. * Provide an archive of the discussions that influenced the contents of the CDR document that we can publish as accompanying material alongside the document itself. ..." Frankly, there is no single good forum for most such discussions at the moment. An increasing number of people find comp.lang.lisp is horrible drain on with decreasing returns, yet it remains the only broad forum in existence -- everything else is specific to an implementation, library, user group, or an application area. Even if a CDR is produced by a separate working group using whatever discussion tools they prefer (eg. coffee breaks) it remains that for third parties to comment on it, they need to either talk directly to authors, or pick a bad forum. Or so unless the author(s) have provided a separate forum for discussing the document -- which has yet to happen for any of the 3 CDRs so far. ,) I naively assume that if a list like cdr-discuss existed, it would get fairly wide readership, and provide a place for such commentary. Its existence would be no onus to the authors that I can see, and by restricting the topic to existing CDRs it should be able to maintain a good signal/noise ratio (which still leaves budding CDR authors without a good place to talk about their plans, but admitting "I'm thinking of a CDR" seems to me like a really bad idea and potentially disastorous to the quality of correspondence.) Take CDR-2 for example: Where would you go to talk about writing more invasive CDR based on it, that would describe an extension to CL that does the same thing, but with the standard GETHASH &co? Where would you go to see if there have _already_ been discussions like that? (Note that these discussions can easily take place long after a CDR has been finalized and the original authors have moved on.) Maybe I'm missing something, but I really don't see how such a list would create barriers for authors -- quite the opposite. Of course it should be made clear that the list has no official standing, but I would hope that some vestigial reading comprehension still lingers on both shores of the Atlantic... > Do you think that cdr-level is appropriate for gmane when it is only > about the CDR process? (That's not a rhetorical question. ;) Appropriate yes (Gmane is a really nice way to read mailing lists you don't want to clutter your inbox), strictly necessary no. Announce list at least would be good to have there, though. > and we also don't want to favor open source approaches over, > say, commercial software vendors. Any specific recommendation could > give the wrong impression here. I'll take your word for it, but I don't really see how saying "we believe licence XYZ to be a good match for the needs of most CDR documents, but you are free to use any licence you want" can be interpreted as favoring open source over commercial vendors -- but clearly I am not the most unbiased person to think about this. In any case, CDR is a damn fine idea, and I believe it will benefit the mythical beast known as the lisp community immensely in the long run. Cheers, -- Nikodemus Schemer: "Buddha is small, clean, and serious." Lispnik: "Buddha is big, has hairy armpits, and laughs." From csr21 at cantab.net Sat Nov 11 12:51:27 2006 From: csr21 at cantab.net (Christophe Rhodes) Date: Sat, 11 Nov 2006 12:51:27 +0000 Subject: [cdr-devel] Re: [cdr-announce] CDR 2 updated In-Reply-To: <36F0BA24-61AA-4D2F-A361-C544E6496D38@p-cos.net> (Pascal Costanza's message of "Sat, 11 Nov 2006 13:17:56 +0100") References: <36F0BA24-61AA-4D2F-A361-C544E6496D38@p-cos.net> Message-ID: <877iy2w3hs.fsf@cantab.net> Pascal Costanza writes: > Ingvar Mattsson has submitted a new version of CDR 2 which is now > available at http://cdr.eurolisp.org/document/2/ I know that this isn't in the original contract, but I'd really like it if previous, superseded versions of CDRs were available, ideally with some description of why a change was made. Would there be scope in the CDR process to either encourage or enforce this, and is it technically feasible to make such earlier versions accessible? (so for example is the same as , while there would also be in this case.) Cheers, Christophe From pc at p-cos.net Sat Nov 11 13:07:51 2006 From: pc at p-cos.net (Pascal Costanza) Date: Sat, 11 Nov 2006 14:07:51 +0100 Subject: [cdr-devel] Re: [cdr-announce] CDR 2 updated In-Reply-To: <877iy2w3hs.fsf@cantab.net> References: <36F0BA24-61AA-4D2F-A361-C544E6496D38@p-cos.net> <877iy2w3hs.fsf@cantab.net> Message-ID: <7E4A779A-9B95-43ED-8CC5-67F719CE3F56@p-cos.net> On 11 Nov 2006, at 13:51, Christophe Rhodes wrote: > Pascal Costanza writes: > >> Ingvar Mattsson has submitted a new version of CDR 2 which is now >> available at http://cdr.eurolisp.org/document/2/ > > I know that this isn't in the original contract, but I'd really like > it if previous, superseded versions of CDRs were available, ideally > with some description of why a change was made. Would there be scope > in the CDR process to either encourage or enforce this, and is it > technically feasible to make such earlier versions accessible? > > (so for example is the > same as , while there would also > be in this case.) I think it should be technically feasible to do this. Currently, the CDR website is done manually, so providing extra directories where old versions are kept shouldn't be problematic. Of course, we can only encourage authors to provide rationales for changes, not force them. The issue that I currently see (and haven't anticipated) is that CDR 0 is fixed by now and cannot be changed anymore (without violating the CDR principles). The text of the CDR process is included in CDR 0, together with what we encourage the authors and submitters to do, etc. On the other hand, CDR itself can start to deviate from CDR 0, and when a sufficient amount of changes have been made to the CDR process, we can issue a new CDR for describing CDR. [1] The recent suggestions by Nikodemus and you aren't strong deviations that are in conflict with CDR's original goals, so I don't see a fundamental problem here. What do the other CDR editors say? Pascal [1] Ah, I love these metacircularity issues.... ;) -- Pascal Costanza, mailto:pc at p-cos.net, http://p-cos.net Vrije Universiteit Brussel, Programming Technology Lab Pleinlaan 2, B-1050 Brussel, Belgium From edi at agharta.de Sat Nov 11 14:29:08 2006 From: edi at agharta.de (Edi Weitz) Date: Sat, 11 Nov 2006 15:29:08 +0100 Subject: [cdr-devel] Re: [cdr-announce] CDR 2 updated In-Reply-To: <7E4A779A-9B95-43ED-8CC5-67F719CE3F56@p-cos.net> (Pascal Costanza's message of "Sat, 11 Nov 2006 14:07:51 +0100") References: <36F0BA24-61AA-4D2F-A361-C544E6496D38@p-cos.net> <877iy2w3hs.fsf@cantab.net> <7E4A779A-9B95-43ED-8CC5-67F719CE3F56@p-cos.net> Message-ID: On Sat, 11 Nov 2006 14:07:51 +0100, Pascal Costanza wrote: > The issue that I currently see (and haven't anticipated) is that CDR > 0 is fixed by now and cannot be changed anymore (without violating > the CDR principles). The text of the CDR process is included in CDR > 0, together with what we encourage the authors and submitters to do, > etc. > > On the other hand, CDR itself can start to deviate from CDR 0, and > when a sufficient amount of changes have been made to the CDR > process, we can issue a new CDR for describing CDR. [1] The recent > suggestions by Nikodemus and you aren't strong deviations that are > in conflict with CDR's original goals, so I don't see a fundamental > problem here. > > What do the other CDR editors say? I think that even if the deviations aren't big, we should nevertheless track them somehow, otherwise it doesn't make sense to have CDR 0 at all. I'd propose that (assuming we want to change something) we either release a new CDR which replaces CDR 0 (which is to be withdrawn) or that we decouple CDR itself from CDR documents. From pc at p-cos.net Sun Nov 19 22:14:11 2006 From: pc at p-cos.net (Pascal Costanza) Date: Sun, 19 Nov 2006 23:14:11 +0100 Subject: [cdr-devel] Some changes to the Common Lisp Document Repository. Message-ID: Hello everybody, Based on the constructive feedback we have got on the CDR process so far, we have made some changes to CDR as follows. + First, due to the fact that the CDR process has been described as CDR 0 which has already been finalized, strictly speaking we aren't allowed to make changes to CDR anymore. We have solved this by resubmitting the CDR process itself as CDR 4. You can find the new version at http://cdr.eurolisp.org/document/4/ together with the changes compared to CDR 0 (which are also described below). Since we expect more suggestions for improvement in the near future, we use an exceptionally long initial period for CDR 4 until November 1, 2007. This allows us to make changes as we go along. At the same time, the CDR process in use will already be CDR 4. This means that the CDR process for submitted CDRs until then will be somewhat unstable, but we think that this gives us the advantage of being able to experiment a little such that the final result will hopefully be a better tested process. Consider this equivalent to the bootstrapping of a metacircular software architecture, or to the development of any new piece of software for that matter, where you have to test and change things as you go as well. ;) + We have added more details about submission of accompanying material, especially with regard to licenses that give us the right to publish it at the CDR website and related publications. We forgot about this issue in CDR 0. + We have added a section where we ask authors and submitters to provide rationales for changes they have made during the initial period. From now on, we also intend to add archives of previous versions of a CDR for better traceability of such changes. + We have added a mailing list cdr-discuss - see http://common- lisp.net/cgi-bin/mailman/listinfo/cdr-discuss for the mailing list page. This mailing list is supposed to be used for discussions about specific CDRs. This is especially useful when CDRs are in their initial period, but may also be useful afterwards. The already existing cdr-devel is purely for discussions about issues related to CDR itself (that go beyond the scope of what is described in CDR 0 and CDR 4). We hope that these changes address the suggestions we have got so far. Best, Pascal -- Pascal Costanza, mailto:pc at p-cos.net, http://p-cos.net Vrije Universiteit Brussel, Programming Technology Lab Pleinlaan 2, B-1050 Brussel, Belgium From pc at p-cos.net Tue Nov 21 20:20:40 2006 From: pc at p-cos.net (Pascal Costanza) Date: Tue, 21 Nov 2006 21:20:40 +0100 Subject: [cdr-devel] Initial period for CDR 3 prolonged. Message-ID: <9A770ACE-BEC1-49E3-A96E-AD0A81F24B28@p-cos.net> Since the cdr-discuss mailing list is only a few days old, and on request by the author, we have prolonged the initial period before finalization of CDR 3 "Revisiting CONCATENATE-SEQUENCE" by Christophe Rhodes until January 15, 2007. Christophe Rhodes welcomes discussions about this CDR at cdr-discuss to which he is subscribed himself. Please mention CDR 3 in the subject line in your postings about this CDR. Pascal -- Pascal Costanza, mailto:pc at p-cos.net, http://p-cos.net Vrije Universiteit Brussel, Programming Technology Lab Pleinlaan 2, B-1050 Brussel, Belgium From pc at p-cos.net Wed Nov 22 10:03:29 2006 From: pc at p-cos.net (Pascal Costanza) Date: Wed, 22 Nov 2006 11:03:29 +0100 Subject: [cdr-devel] Initial period for CDR 2 prolonged. Message-ID: <1A0FE98D-EB85-4777-A041-76525C8BD44B@p-cos.net> Since the cdr-discuss mailing list is only a few days old, and on request by the author, we have prolonged the initial period before finalization of CDR 2 "A generic hash table interface specification for Common Lisp" by Ingvar Mattsson until January 15, 2007. Ingvar Mattsson welcomes discussions about this CDR at cdr-discuss to which he is subscribed himself. Please mention CDR 2 in the subject line in your postings about this CDR. Pascal -- Pascal Costanza, mailto:pc at p-cos.net, http://p-cos.net Vrije Universiteit Brussel, Programming Technology Lab Pleinlaan 2, B-1050 Brussel, Belgium From pc at p-cos.net Wed Nov 22 10:14:56 2006 From: pc at p-cos.net (Pascal Costanza) Date: Wed, 22 Nov 2006 11:14:56 +0100 Subject: [cdr-devel] CDR announcements / mailing lists Message-ID: Hi everyone, I have posted the recent two announcements about prolongation of CDRs to all three CDR-related mailing lists. We don't have the intention to post such and similar announcements to all three lists in the future. I have only done this because I assume that currently, people may not be signed up to the appropriate mailing lists, especially because cdr-discuss has only been a recent addition. In the future, we will restrict announcements to cdr-announce exclusively. As advertised, cdr-discuss is for discussion about specific CDRs, and cdr-devel is for discussion about CDR itself. There may be some overlap between cdr-discuss and cdr-devel when it comes to the CDR document describing the CDR process, but I guess this will pose no serious problems. Cheers, Pascal -- Pascal Costanza, mailto:pc at p-cos.net, http://p-cos.net Vrije Universiteit Brussel, Programming Technology Lab Pleinlaan 2, B-1050 Brussel, Belgium