Path: utzoo!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!uunet!usenix!jsq From: jsq@usenix.org (John S. Quarterman) Newsgroups: comp.std.unix Subject: Re: Standards Update Poll (comments in responses) Message-ID: <507@usenix.ORG> Date: 12 Sep 90 19:23:30 GMT References: <440@usenix.ORG> <504@usenix.ORG> <505@usenix.ORG> <506@usenix.ORG> Sender: jsq@usenix.ORG Lines: 746 Approved: jsq@usenix.org (Moderator, John Quarterman) X-Submissions: std-unix@uunet.uu.net From: jsq@usenix.org (John S. Quarterman) 8a: What other committees should be covered? response 6 I'd like reports on ISO committees (I'd *love* to hear snitches about the ISO Prolog committee). response 8 I'm not sure what other committees exist :-) response 10 I tried. response 11 I believe that Usenix should not be particulary interested in standards. Let Uniforum or other bodies do that. (If you were as large as IEEE I could see it) I would rather see the emphasis of Usenix on OS research and development. I have enjoyed recent papers on distributed systems and such and have less interest in standards and such. One thing that I really like is Usenix's sponsorship of scholarships for students and such. One thing I really hate is Usenix's sponsorship of things like Uunet or sponsoring the development of public domain uucp's and the like. response 12 I run an academic computer center. The information I get now enables me to plan better for the future, and sometimes to write to someone if I see something that looks particularly good or bad. response 16 I think it would be great if you could provide an overall view (once) of what each group is trying to accomplish, details on a subset of the groups, and a "floating" review that moves through some of the less popular groups covering, for instance, one per month. response 17 Make sure that each POSIX committee is covered. Cover networking standardizationbeyond 1003, i.e. 802. response 22 [Personal view only!] Email and Enews are a highly efficient way of covering, tracking, and operating the standards process which must include - -identification of standards-needs -debate of technical and commercial issues in the decision of work on a standard -identification(if possible) of an existing de facto basis for a de jure standard -discussion of technical and commercial issues in formulation of a standard -circulation of drafts, contributions, etc -circulation of suggested modifications, arguments, etc -voting The USENIX participation in Enews and Email forms a valuable informative contribution. It could be extend to promote some or all of the above between its members and amongst other standards-related workers response 24 We'd like to see a regular report on the supercomputing committee; only thing we've seen so far was a paper at the April 90 CUG meeting. response 36 P1201.* The ISO JTC committee on icons, etc. response 37 Few: committees that are working on well-legitimized standards subjects (e.g., 1003.1, .2, but not .4) should be covered well. Less legitimized standards subjects should be mentioned and documented, but there's already enough heat and light emanating from them that we don't need any more coverage. response 40 X3V1 for printing standards, ODP Distributed Applications work, P1203 User Interace work. response 43 I like the snitch reports. I think that some of my answers may be misleading. For example, I said that I do not read the snitch reports in ;Login. That is true because I have already read them on comp.std.unix. It does not mean that I am not interested. response 45 Usenix is the only brake I have found on the Standards Steamroller. We need better, more elegant standards, in the tradition of Unix and TCPIP and fewer monstrosities like X and OSI. response 50 The Mass Storage Standards Committee should be covered. response 51 The uncovered TCOS groups and X3J16 (I'm working on it). response 61 Interface standards and Languages response 64 The ones currently covered are the only ones I know, so how can I answer this question? response 67 Not familiar with full extent of current coverage, but am interested in SGML and other document-oriented standards (eg, the initiative sponsored by Assoc. Comp. Linguistics et al.); this may or not be of interest to Unixers in general response 68 Interesting effort. I must confess that I answered 3, because in many cases I don't KNOW what you are currently doing. We (sun) have lots of internal traffic about standards efforts, and I don't personally follow yours other than via the newsgroup. One only has a finite amount of time.... response 70 Keep up the comp.std.unix POSIX.* snitch reports. Try to have them follow the meetings by no more than a month. response 75 | response 76 The X/Open work and their effect on POSIX and vice versa. More on ISO POSIX. response 79 My professional interest and an area of vital importance to the future of UNIX as it becomes more distributed via RPCs and such is high speed networking.. at a minimum things like XTP over FDDI, HIPPI esp the datagram work, SONET. The SW like groups I would be most interested in following are the POSIX threads people and the RPC people (I think there is some such working group), but we have been mostly involved at the HW level to date and I have just done a cursory reading over comp.std.unix. ..... I do think there is a potential for too many, too undefined standards and would urge your group to be careful. IMHO the whole OSI mess shows the danger of too many cooks. The thing that most offends myself (and my boss) is that you can't just anon FTP copies of OSI and such like standards from the NIC. We actually bought paper copies of a few we thought might be relevant. When we got them they were: expensive, lousy xerox copies, out of date. But what do I know anyway, I do hardware. response 88 Add non-POSIX committees (e.g. X3) which have impact on UNIX, C, etc. response 91 This is a very difficult question (as I'm sure you know). You can't cover everything with limited resources, yet there are many standards bodies which are having an effect on (yechh) Open Systems. Perhaps a coordinating and synthesis role is more appropriate for user groups. For example, how many UNIX users know about the intersecting effects of TCSEC, OSI, NIST anmd other bodies on UNIX contents and interfaces? I guess as many committees as possible with reasonable quality... response 92 The problems I have with the standards committees and covering them is that I get the feeling the "common user" is not invited. While it is necessary to hear from the industry gurus and vendors, I have a feeling all this is going over the heads and behind the backs of those of us in the trenches who will have to work with these standards later. There has to be some way to include the users in the process. And that's the problem. I would have liked to be involved with the ANSI C standards committee and some of the POSIX committees but either I didn't find out about possibly getting involved until too late or I don't have the time of the executive of a software house to pursue membership. Avenues for "part-time" members should be more open then they have been and allowed to be filled by different people. Additionally, there should be a better distribution method for documents reguarding the standards. By the time I've seen some of these documents, they've gone through another set of revisions and when I comment on them, I sound like a fool because the concerns were already addressed. If I can get involved in a standards committe, I would. I just can't make it a full time effort but would be willing to do the best job I could with the time I can put into it. response 96 Language committees if they relate to UNIX (Fortran, perhaps). 8b: What committees should *not* be covered? response 16 All groups should get some coverage. response 37 See previous comment -- let's not spend USENIX resources on the set of these activities that are out of control. Let's simply point out that these exist and are controversial and let those who are interested find out more about the controversy. response 40 COBOL, Fortran response 43 I don't care much about eurpoean standards which are not world standards. If fact, if your coverage were limited to American National and ISO standards, I would be happy. response 45 Usenix has limited resources. We should not dilute the coverage to the point that the Usenix influence ceases to be felt. response 51 I don't think it makes sense to cover groups that are largely done, like the C standards group. Having said that, I think that there's still a lot of interest in groups like 1003.1, that should be done but aren't. response 64 The ones that are currently covered are fine - I do not reccommend dropping any response 75 | response 88 Continue current coverage, plus above. response 92 All should be covered. Including hardware standards (i.e. bus). response 96 I don't think much of the OSF and UI, but they're going to have an effect so I guess I'd like to be informed of what they're doing. 8c: What else should the USENIX Standards Watchdog Committee do? response 16 Provide an overall view for Usenix, subsidize costs for Usenix members to take a more active role. response 24 The Committee should lobby the appropriate standards committees on issues they feel are of significant importance. response 30 Keep the standards bozo's from screwing everything up. Thanks for the white paper on sysadmin standardization. response 51 I'd like to see USENIX continue influencing standards. I think that it can best do so by sponsoring thoughtfully written pieces of various sorts, and by active collaboration with other users' groups. response 54 Always be aware of standard practice and the effects of new initiatives. It does no good to specify an interface that will break a significant number of existing applications. response 58 So long as you're letting the membership know what is going on, ina timely manner, that's about all that you need to do. response 64 Nothing more or less than it does, provided that it is able to cover all the committees response 69 Produce a dynamic "summary" document to allow "users" to know the current status of various efforts. Include as attachments drafts and standards and provide updates as needed. Also address FAR's FIPS etc. for government users. Charge for this service as needed to break even. response 75 | response 79 Keep an eye on those folks at NIST! response 88 Leverage current activities through cooperative ventures with other major user groups or associations. response 91 See answer to 8a response 92 Provide a louder voice for the programmer in the trenches and the forum or the entry to voice those opinions and have them taken seriously. Or at least until the explanation as to why the idea will not work. response 96 Send feedback into the committees. 8d: What should the USENIX Standards Watchdog Committee *not* do? response 16 All the work themselves! ;-) I would like to see more general participation. response 37 It should not submit positions, nor forumlate positions, on standards. response 40 Stay out of marketing turf battles like UI vs. OSF. response 43 I don't believe that the Watchdog Committee should turn into a barking or biting dog. Just report. Do not become part of the process. Do not become enmeshed in the politics. response 51 The committee should not turn into a formal bureaucracy. I like the volunteerism it trumpets. For me, it is a reminder that people--not corporations--are still the key to UNIX. response 54 Never strike an alliance with a vendor or vendor consortium unless that consortium has a record of fair play. response 58 Assume that it knows all the right answers, or that it should be the only source of information available, or that for some reason it is necessarily the best source of information (although maintaining that as a goal would be a nice touch) response 75 | response 79 Generate new standards response 88 Be flippant about the process of consensus building. response 91 Another vexed question. Whether user groups should form into lobby groups for standards activity is difficult - I'm aware of the "world standards" initiative, and I think that it's worthwhile. It's also enormously politically difficult, of course :-) response 96 I have no complaints with what they've been doing so far. It should be obvious that too much input from vendors is a dangerous thing, so I'll just leave it at that... 8e: What else should USENIX do regarding standards? response 6 Contribute to the criticism of *existing* standards. Reports on the effect that existing standards have had, the extent to which they are observed. Ok, it's not feasible to do a lot of this, but it would be useful to know say, how much attention I should pay to ASN.1. Particularly when there is a continuing "interpretation" process, as for Ada and C, it would be nice to hear about those things. response 9 Promote reference implementations of standards. The X Window System is one example; there should be others. For example, you could modify GNU utilities to produce reference implementations. response 16 I think that Usenix should take a more active role in the standards areas. I personally would be interested in particpating on some of the reviews. response 17 Workshops. response 21 Discourage standardization of immature technology. response 24 I'd like to be able to get an update on FIPs activity from comp.std.unix. I have all the names and numbers to call at NIST, and they are very helpful there, but when I have a question about the status of a FIPs I figure a lot of other people probably do, too, and why not answer all of us at once in a public forum? response 39 Lobby to maintain online (electronically accessible) copies of software standards. Yes, I know that sales and publication provide the income which allows the standards committees to go on creating standards, but if you ask me, there could stand to be a bit less of that in the computer software arena anyhow. Although actually, I think having electronically accessible standards documents (and drafts, especially) will, if anything, increase interest in the standards, and the number of potential participants. response 40 USENIX should take a look at the standards process and its value to its members. This should be done by a special committee of the BoD. In addition to providing valuable information, such a study could help guide BoD decisions. response 43 It would be great if current drafts were available from uunet. I know that the standards organizations need to generate $ by selling standards, however, they charge rip-off prices. Global Engineering wanted $75 for a draft of X3.159. The final standard *only* cost $40 with my ANSI member discount. [[BTW -- My company contributes over $50,000/year to ANSI]]. --- The main reason that I want the documents on-line is for ease of access and not for cost savings. I know Hal generates postscript as part of the document generation process. The postscript files could be made available. That would not expose the troff source to the world. response 50 Take an active role in getting the information out. Why aren't white papers and committee minutes on-line? You might get more involvement if people could ftp information from some place and read it. response 51 Anything to support users' work to advance UNIX. response 54 USENIX needs to be active in ISO and IEEE committees to protect the interests of users. The visibility of modern-day standards efforts has attracted hundreds of vendor representatives who are struggling to take control of various focus groups. response 58 I'd tend to think that given that we have a group reporting to the membership about what's going on in committee, that there should be some way to solicit input from the membership about the material reported and feed that back into the standards process. response 61 Hmmm. Sometimes I think too many diverse interestes are doing too much. But when the good folk need support on SC22 for some dumbo's proposal, we need all the help we can get. And no, you can't quote me on that. response 75 | response 77 Just keep involved please.... response 79 One thing that seems to be missing is a database on what is available that complies to std umpty ump, whether it has passed conformance test XXX, if it has know problems working with vendor Z's also conforming umpty ump product. Maybe there is on opportunity here. response 88 Coordinat ballots with other institutional reps response 92 Be a more visable presence. response 96 Encourage extensions and alternatives. There are things being standardised that are way premature: system administration, for one, or windowing. I think building standards from nothing, or standardising on a clearly clumsy technology (X) is worse than no standards at all. The System V.3 system administration suite is the best I have seen on an actualy working UNIX system, and should be given quite a bit of weight... it's the only existing practice worth a damn. If someone could put pressure on Sun to dedicate NeWS to the public domain it would save Sun's and everyone else's bacon... 8f: What should USENIX *not* do regarding standards? response 16 It is important that Usenix not get itself dragged into the middle of all the standards activites and not get into the "poltics" of the activites more than it has to. It can provide a good "non-aligned" and technical view. response 29 Have any of its own, there's too many competing outfits as it is response 37 See previous comment -- it should not take positions. response 43 Don't take technical positions. Each of the members is capable of expressing himself. response 51 I don't think it makes sense for USENIX to duplicate the efforts of UniForum. The UniForum technical committees and the POSIX Explored documents are praiseworthy; we should encourage, but not imitate them. response 58 Try to set itself up as the governing body for standards creation, or as the "owner" of any of the standards. response 75 | response 77 Support the opinions of individuals, i.e. especially board members, to the standards committees. Try only to do the best at supporting the best interests of *ALL* members. response 81 Do not ignore the standards. response 92 Sit in the background and only watch. response 96 First, do no harm. Don't get caught up in the standards bandwagon: don't get behind standards for the sake of standardising. Some things aren't ready. 8g: What else do you want us to know? response 5 With my not-so-perfect English language knowledge, I had some difficulty in understanding some questions (they being so brief and not too explatonary), so it might be that my answers do not really represent my opinions. response 6 For a lot of the questions above, I didn't really mean "3", what I really meant was "don't know" or "don't care". response 8 Basically I'm happy with what is now going on. response 9 You should consider collaborating with the League for Programming Freedom regarding current attempts to copyright and/or patent software interfaces. Such attempts are in direct conflict with standard setting, and will gravely hurt the software industry in the future. response 13 As you can probably tell from my answers, I tend to ignore the standards process. Thus, I don't have strong opinions on how the process should be done or changed. However, I am glad that someone is paying attention, and I like the reports that keep me apprised of what is happening. response 16 It is good to see the coverage of the standards in the first place. I think a lot of technical people have been left out, because they didn't know how or what to do. response 18 I don't really care about most of this, but your poll didn't give me an option to indicate that. Therefore, some of the answers you got for the above are meaningless. Basically, I think standards are mostly a good thing, and I'm glad some people are interested in them, and if I ever want to get involved I want to know where to go. In the meantime, I really am not interested in seeing extensive reporting on the issues. Question 7 left our "educator" and "researcher" -- I'm both. Sun Aug 19 22:26:48 EST 1990 response 22 The above is purely a personal view and does not necessarily represnt the view of Data Logic or any of its clients response 24 I find the electronic mailing list, the snitch reports, and the regular summaries on Standards, Groups, Publications, and Meetings invaluable and would hate to see them stopped or curtailed. Before you do that, please tell us what it would cost to keep them going. response 26 response 34 - The on-line standards reports have been invaluable to me. They are excellent. (I work in the Oak Ridge National Laboratory Office of Laboratory Computing, responsible for computing policy and future directions.) response 39 Since I get ;login: and occasionally read comp.std.unix, it would be nice if the reports were more clearly labeled by date of writing, or number or something. I sometimes end up reading the same reports more than once as a result. Also, a bit more editing of the reports wouldn't hurt - there's an unfortunate tendency towards long-windedness. Finally, the standards reports seem to have taken over ;login. I know that a lot of the more academic articles are now in Computing systems, but I miss the more frequent, rough-edged but thoughtful or useful articles that used to be in there. There are at least some of us who are still hoping that not all research goes on within (or in the context of) standards committees. I guess that it would be a good idea to split off the standards reports into a separate newsletter (though I probably wouldn't pay extra money for it). Perhaps limiting them to quarterly issues (or less!) might be enough. response 42 I answered "3" to a bunch of questions to indicate "no opinion" since this program didn't let me just leave a question unanswered. There are plenty of subjects which I don't have any idea how much usenix is doing now, so I don't have an opinion on more or less (for example). response 43 You had a list of questions about publications and user groups. Some of these I never heard of. I don't recall them from the publication lists on comp.std.unix. Maybe you could update those lists. response 45 Whatever happens, please don't REPLACE the newsgroups -- augment them. response 46 I have been planning on joining Usenix. I would rather read these reports in the news group than in ;login dues to timeliness. Note you have a bug in your survey (2 5H questions). response 48 One area I would like to see more standard is the Addressing of Email. I dislike uucp only sites being second hand citizens. response 51 I'll kick myself later for letting this straight line pass. response 54 POSIX committees appear to be considering UI/OSF politics in some of their actions and that is wrong. Let's keep in mind who we are trying to protect: the end-user and the application developer. Let's lobby POSIX to adopt standard practice, to standardize only those areas in which there is demand for standardization, and to always hold their meetings in areas where there is a large concentration of *users.* response 60 I basically just browse the standards report in ;login: and on-line (mostly in ;login:). I mostly have no opinion regarding these questions. response 64 I appreciate the importance of standards, but it's all too easy to get lost in the multiplicity of committees. response 65 I like the context provided by the reports, but I usually get confused by (1) the proliferation of standards groups, many of which seem to have overlapping charters; (2) the alphabet, er, number soup game ("let's see, .1, that's, uh, system calls?"). It would be good if this could be clarified every now and then, but it's probably not worth doing in every issue of ;login. response 71 Generally happy with current state of affairs; is not broken and does not especially need fixing, from my perspective. (Well, except for excessive enthusiasm for long tedious polls... :-)) response 75 | | response 76 I am also a member of EUUG-S (European UNIX User Group in Sweden). response 77 You've all done very well so far. Keep up the good work. I really like this poll, and the simple way in which it works. response 78 Now that I'm no longer on a P1003 working group, the ;login, snitch reports, etc are great ways to keep in touch with Posix land response 81 I am appalled that, despite being POSIX conformant (or nearly so?) BSD UNIX -- vastly easier to use -- is so little represented in commercial UNIX products. Furthermore, references to USENIX appear almost never in the commercial press. Both USENIX and BSD UNIX have a whole lot to offer commercial business, and I'd like to see them as widely known as they are valuable. response 84 I have not had much experience with standards forming commitees, hence the lack of expression of strong opinions above. I do not have a lot of spare time to devote to keeping up with evolving standards but I have found ;login: 's coverage informative. I've occasionally read some standards reports in UNIX review but cannot at this time justify a subscription - hence the 'n' reply above. Coverage of the general directions of evolving standards is all I really need and ;login: satisfies that fairly well. Technical detail is really only needed by me to understand certain controversies (i.e. clarification of the 14 character filename limit in POSIX 1003.1 WRT BSD and Sys Vr4). response 86 Are you interested in doing more about any other issues regarding UNIX aside from "standards"....seems to me there are some general philosophical issues that will be affecting UNIX i.e. Lotus court case...that might justify some involvement by USENIX... response 87 The editor's plans outlined in last ;login: seemed good. response 88 Each of the current user groups/associations tend to represent distinct segments of the user population. There are, however, significant overlaps of activities. Better cooperation between groups and associations, such as cooperative ventures on standards activities, would go a long way toward improving the UNIX community and showing a more united front to those organizations which are migrating to UNIX/open systems. Having UNIX-Democrats and UNIX-Republicans is OK (read GOOD THING), but having each functioning in an insular manner is not (read BAD THING). response 92 I want to know how to get involved even on a part time basis. I reallyy thing there's a body of knowledge and insight being lost by not contacting those of use with limited time. response 93 Although i only occassionally get through enough net news to reach this newsgroup, i will attempt to do so more frequently now that i've discovered you all produce these reports on standards. I would hope these public contributions will not be discontinued. Thanks! response 95 Well, the next time you make such a poll, you might consider leaving an option to *not* answer a question in your script. To a lot of the questions, I simply do not have any good answer. As it is, I could only guess as to a neutral one... response 96 I think y'all are doing a great job. Keep it up. Volume-Number: Volume 21, Number 101