jsh@usenix.org (Jeffrey S. Haemer) (12/03/89)
From: Jeffrey S. Haemer <jsh@usenix.org>
An Update on UNIX* and C Standards Activities
December 1989
USENIX Standards Watchdog Committee
Jeffrey S. Haemer, Report Editor
IEEE 1201: User Interface Update
Eileen Coons <coons@osf.org> reports on the October 16-19, 1989
meeting in Brussels, Belgium:
"The time has come," the walrus said,
"To talk of many things:
Of shoes -- and ships -- and sealing wax --
Of cabbages -- and kings --
And why the sea is boiling hot --
And whether pigs have wings."
-- Lewis Carroll
The P1201 committee is on a divine mission to define standards for
user interface technologies. Lewis Carroll would have loved P1201
meetings.
In keeping with the precedent set by previous P1201 meetings, this
latest get-together was spirited. The quasi-good news is that, by the
end of the session, not one, but 3 PAR's had been defined, as the
group split into 1201.1 (Application Programming Interface), 1201.2
(Drivability - Look & Feel), and 1201.3 (User Interface Definition
Language). One participant aptly named the proceedings "PAR Wars".
There was agonized discussion over the various sub-group's missions,
and an equal amount of agonized, and at times agonizing, wordsmithing
over the .1 and .2 PAR's themselves. The .3 group thoughtfully
elected to split off and define itself in private. The PAR's will be
submitted via proper official channels to be blessed at the January
SEC meeting.
For anyone not familiar with the PAR process, PAR is an acronym for
Project Authorization Request. An individual or group that believes
some work should be done by an IEEE committee drafts a document
describing the work, which is then submitted to the IEEE as a PAR.
__________
* UNIX is a registered trademark of AT&T in the U.S. and other
countries.
December 1989 Standards Update IEEE 1201: User Interface
- 2 -
Usually the PAR is circulated to the IEEE membership in one of its
mailings.
The SEC (Steering Executive Committee) reviews the PAR during its next
scheduled session, typically held during a POSIX meeting. The SEC
votes on the PAR, and if the PAR is approved by the SEC, it is
presented to TCOS (Technical Committee on Operating Systems). TCOS
decides in which committee the work will be done. In the case of the
PAR for User Interface, TCOS elected to divorce the work from the core
POSIX effort (1003), and created P1201.
The PAR becomes part of the statement of work and basic charter for
the group doing the work.
Fortunately, at this meeting the group finally created some real
structure for itself. Ninety minutes into the meeting, the group
decided to define an agenda! It also resolved that all meeting
attendees should receive minutes of the meeting, e-mail snafus
notwithstanding. Jim Isaak, the chair of the 1003 SEC, helped with
structural definition by supplying IEEE rules and charter information,
explaining the balloting process, and listing action options open to
the committee.
Seven ballot alternatives were proposed, ranging from submitting a
proposal for immediate ballot, to disbanding 1201, packing our tents,
and going home. A vote was called, and although there was no
consensus (hardly a surprise), the heavy favorite was a proposal to
adopt Motif's API as the basis for a standard API specification, and
to extend it to accommodate aspects of Open Look's look & feel.
This general direction was unpopular with a vocal minority, however,
so the group took a break then reconvened, discarded the vote and
returned to its original, pre-poll path of action: defining a
specification for an API based on neither Motif nor Open Look, but on
some new API -- probably a hybrid of the two.
[Editor's note: I've heard more than one person express ill-ease about
the restricted range of choices being considered. Why is there no
mention of NeXT/Step, for example? A noticeable feeling among people
who aren't on the committee is that it's too early to try to
standardize in this area, and that the answer to the question, "Motif
or Open Look?" should be, "No thanks."
The answer to the implied question, "Why is there a P1201 and why are
we doing this now, anyway?" seems to be is that NIST, the National
Institute for Standards and Technology (the people who bring you
FIPS), is pushing hard for rapid creation of a GUI standard.]
Two presentations were made: one by AT&T, in favor of the joint API
concept, and one by OSF, arguing against its feasibility. In an
unusual and unfortunate departure from Robert's Rules of Order, this
December 1989 Standards Update IEEE 1201: User Interface
- 3 -
was followed by a critique of -- some thought, attack on -- the second
presentation by one of the acting chairs, Clive Feather of X/OPEN.
P1201 may be many things but, so far, staid isn't one of them...
On a more neutral note, several representatives from organizations
working on UIDL technologies made presentations about what they were
doing in that arena, and then went off to form P1201.3. God bless
them.
The rest of the group broke into the .1 and .2 sub-groups for working
sessions during most of the remaining meeting time. Each group
reviewed its newly drafted PAR. P1201.1 also spent time comparing
Motif and Open Look, identifying and exploring the differences between
the two API's, and looking for potential drivability issues that could
be deferred to P1003.2. P1003.2 took a similar course of action,
comparing the looks and feels of the two technologies.
It's rumored that the .1 group will be meeting Dec. 4 - 5 in
Cambridge, MA to pursue their quest for a merged API. Interested
parties should contact Betty Dall, AT&T, for more details. (E-mail
ejd@attunix.att.com, or phone Betty at 201-522-6386.)
There was also a spirited discussion regarding when and where the next
P1201 meetings should be held. After various alternatives were
explored, and only two (or was it three...?) votes, the group decided
to keep P1201 meetings in the same vicinity and timeframe as POSIX
meetings, since many attendees need, or want, to participate in POSIX
as well.
All in all, it wasn't too bad. The weather in Brussels was nice, the
Belgian beer was pretty good, and the meeting was, um...,
entertaining.
December 1989 Standards Update IEEE 1201: User Interface
Volume-Number: Volume 17, Number 83jsh@LONGWAY.TIC.COM (05/03/90)
From: <jsh@usenix.org>
An Update on UNIX* and C Standards Activities
January 1990
USENIX Standards Watchdog Committee
Jeffrey S. Haemer, Report Editor
IEEE 1201: User Interface Update
Peter H. Salus <peter@uunet.uu.net> reports on the January 8-12, 1990
meeting in New Orleans, LA:
What's happening?
P1201 purports to concern itself with the user interface. As of the
New Orleans meeting, P1201 comprised .1 (Applications Programming
Interface), .2 (Graphical User Interface), .3 (Human-Computer
Interaction), and .4 (XLib) subgroups.
Working backwards through these, 1201 has recommended that XLib go to
ballot directly, a proposal which seems to have so shocked the SEC
that they put off deciding on balloting till April. Steve Jobs told
the USENIX audience in Phoenix, in June 1987, that X was ``brain-
damaged''. Whether that's true or not, X has won, and just putting
XLib to a vote makes good sense.
1201.3, under the chairmanship of Richard Seacord, has had a number of
interesting discussions and presentations (of which I attended
several, though not all). The major problem here is that we are
nowhere near knowing what the ``standard'' for an interface might
really require. However, the explorations are valuable, and a forum
like this can be informative.
This leaves me with the GUI and the API. Both in Brussels and in New
Orleans were skirmishes in the GUI wars: battalions of employees of
OSF its member companies arrayed in opposition to those of UI or USO
and theirs, with a pair of observers from NeXT and Apple taking and
placing bets on the sidelines.
I assure readers that have never attended these meetings, acrimonious
backbiting and vituperation are the order of the day in both camps.
Though a former employee of OSF, I wouldn't hesitate to condemn the
behavior of both sides, but the blame rests elsewhere. Where? In the
tourists. See below, but for my money, too many folks like to travel
and too many people have caught the ``open systems/open standards'' bug.
__________
* UNIX is a registered trademark of AT&T in the U.S. and other
countries.
January 1990 Standards Update IEEE 1201: User Interface
- 2 -
So long as the market remains unsettled about Motif, NeXTStep, OPEN
LOOK, and Presentation Manager (to say nothing of Apple's MacIntosh
interface and IBM's CUA) [Editor: That's ``Common User Application'',
a part of SAA.], the meetings of 1201.1 and 1201.2 will serve as
tilting grounds, not occasions for useful discussion.
>From my point of view, until the market (which means the big boys and
the users) has a shake-out, .1 and .2 can only serve as debate
platforms or end up recommending standards that are either the
intersection of OPEN LOOK and Motif or their union. It might be that
2 can come to some sort of conclusion on the various style guides
without .1, but I see the products being waved, not the function
banners.
Why is it turning out this way?
All of this is prologue (``The past is prologue,'' writes Shakespeare
in The Tempest) to a commentary on the TCOS-standards industry.
[Editor: TCOS, the Technical Committee on Operating Systems, is the
IEEE organization under which both 1201 and 1003 fall.]
Over the past 40 years, ISO has approved or accepted over 20,000
standards, which concern almost everything imaginable from hockey
masks to medical prostheses to the hinging of radar masts on inland-
waterway vessels. The standards have arisen in a variety of ways,
most emanating from one of the regional or 70-odd national standards
bodies. Typically, it has taken from four to ten years to progress
from raising a committee to approving a standard. The result of this
has been general agreement within the concerned industry prior to the
issuance of an international standard. Wall plugs are an excellent
example of what happens when the engineers and bureaucrats issue a
standard without industry consensus.
I am far from convinced that the ever-increasing number of 1003 and
1201 (sub)committees is productive or useful, and embarrassed and
appalled at their continuing proliferation. There are currently at
least six or seven standards for diskettes. Do we really need that
many for graphical user interfaces? I think not. Might we get what
happened in the record industry (i.e., 45s for short cuts; 33s for
long works and anthologies) if we wait? I think so.
Moreover, does the standards process really require more than two or
three folks per company? There were 38 in attendance at the ISO/IEC
Joint Technical Committee on Application Portability meeting in
September (including the secretariat); there were nearly 300 in New
Orleans. My perception is that going to a POSIX meeting is a perk.
Holding the meetings in Hawaii, New Orleans, and Snowbird does little
to dissuade me. The New Orleans host was OSF; the Snowbird host is
Unisys. Though the new Unisys is a big entity, I didn't realize they
had a site in Snowbird; nor OSF one in New Orleans.
January 1990 Standards Update IEEE 1201: User Interface
- 3 -
C'mon, lets get back to work, not meetings for the holiday or for the
sake of meetings. 1003.1 did good, solid work. Some of the other
groups are doing work, too. Partying ain't part of it. Bah!
January 1990 Standards Update IEEE 1201: User Interface
Volume-Number: Volume 19, Number 34"Moderator, John S. Quarterman" <std-unix@LONGWAY.TIC.COM> (05/03/90)
From: Jason Zions <uunet!cnd.hp.com!jason> I couldn't let Peter Salus' report go without comments. > ... 1201 has recommended that XLib go to >ballot directly, a proposal which seems to have so shocked the SEC >that they put off deciding on balloting till April. Steve Jobs told >the USENIX audience in Phoenix, in June 1987, that X was ``brain- >damaged''. Whether that's true or not, X has won, and just putting >XLib to a vote makes good sense. Peter leaves out some important details which make the SEC action appear somewhat more intelligent. The primary issue raised related to exactly which specification of XLib was to become the standard. In other words, whose document would get the IEEE document number? The MIT Xlib spec? Which one - X11R3 or R4? Are there changes for R5? Is the document technically correct? What about X/Open's version of the Xlib spec - is it cleaner? Tighter? Easier to understand? More accurate? Is there a specification of Xlib detailed enough to permit implementation of a new interoperable version? The SEC didn't delay specifically to April; they delayed action until the PAR sponsors could develop adequate answers to these questions. >Over the past 40 years, ISO has approved or accepted over 20,000 >standards, which concern almost everything imaginable from hockey >masks to medical prostheses to the hinging of radar masts on inland- >waterway vessels. The standards have arisen in a variety of ways, >most emanating from one of the regional or 70-odd national standards >bodies. Typically, it has taken from four to ten years to progress >from raising a committee to approving a standard. The result of this >has been general agreement within the concerned industry prior to the >issuance of an international standard. Wall plugs are an excellent >example of what happens when the engineers and bureaucrats issue a >standard without industry consensus. I think you'll find there is no ISO standard for wall plugs. Every country for itself, and some take several. (We all know that, when one buys an appliance in the U.K., one must also buy a plug for the end of the power cord and install it oneself or with the help of one's electrician...) >Moreover, does the standards process really require more than two or >three folks per company? There were 38 in attendance at the ISO/IEC >Joint Technical Committee on Application Portability meeting in >September (including the secretariat); there were nearly 300 in New >Orleans. My perception is that going to a POSIX meeting is a perk. >Holding the meetings in Hawaii, New Orleans, and Snowbird does little >to dissuade me. The New Orleans host was OSF; the Snowbird host is >Unisys. Though the new Unisys is a big entity, I didn't realize they >had a site in Snowbird; nor OSF one in New Orleans. The opening sentence of this paragraph seems to be a non-sequitor with respect to POSIX, not to mention the rest of the paragraph. Membership in a POSIX working group or ballot group is independent of one's employment affiliation; each person is accredited as a bona fide technical expert. More than that, many companies do indeed send only one or two people to the meetings. Larger companies may send one person to each committee. If all the standards in development may affect the course of business for a vendor, why should that vendor *not* actively participate in the development of those standards? It may indeed be going overboard for a company to pay for more than one employee to attend a single committee, but even that's not true in all cases; in the case of 1003.1, an HP employee chairs the group and hence cannot really pursue any particular corporate agenda; for HP's views to be represented, an additional person needs to be there. I fail to understand your objection to active participation in voluntary standards making. Why should only three or five people meet in a room and develop a particular standard? If it takes 30-50 people an extra year to develop a better standard, or at least one with wider concensus and greater industry buy-in, what's the problem? Finally, regarding the matter of meeting venue. Unisys is headquartered in Salt Lake City. You tell me - where are the largest meeting facilities likely to be? Where can one obtain low-cost meeting facilities at the end of April in Utah? Were you unhappy with the New Orleans venue? Was the hotel price exhorbitant (given the number of meeting rooms required)? Where would you have preferred we had met, given the constraints of price, air-travel connectivity, number of hotel rooms needed, and number of meeting rooms needed? >C'mon, lets get back to work, not meetings for the holiday or for the >sake of meetings. 1003.1 did good, solid work. Some of the other >groups are doing work, too. Partying ain't part of it. Bah! You're quite right. Partying is not relevant to the Monday-Friday 9-6 work of the meeting. If you see working groups goofing off during the week, feel free to name names and point fingers. Tarring all 1003 groups save 1003.1 (past-tense, as well!) with the same brush of laziness is unfair (not to mention terrible reportorial practice). And yes, having the Sunday before and the Saturday after a meeting in a pleasant locale *is* a perq for many of us. Most attendees work damn hard during the course of the week. The meetings have to be help *someplace*; if the cost can be maintained at a reasonable level, why object to a nice location? Jason Zions Chairman of, but not speaking for, 1003.8 POSIX TFA Volume-Number: Volume 19, Number 36
stanh@meyerhof.bcm.tmc.edu (Stan Hanks) (05/04/90)
From: Stan Hanks <stanh@meyerhof.bcm.tmc.edu> "Committees are, by nature, timid. They are based on the premise of saftey in numbers; content to survive inconspicuously, rather than take risks and move independently ahead. Without independence, without the freedom for new ideas to be tried, to fail, and ultimately succeed, the world will not move ahead but live in fear of its own potential." -- Dr. Ing. h.c. F. A. Porsche, a long time ago For years I have contended that the only standard worth a damn was one which was a codification of existing practice rather than one which was formulated by a room full of people who have a vested financial interest in the way the standard comes out. Peter Salus is absolutely right. It's time to stop this shilly-shallying about with standard this and standard that, and let us get back to doing useful work. We'll talk about standards (particularly in the user interface area) later -- when we actually have something to talk about (what a novel idea!). Stanley P. Hanks Director, Information Technology Planning and Development Baylor College of Medicine, One Baylor Plaza, Houston TX 77030, Mail Stop: IR-3 e-mail: stanh@bcm.tmc.edu voice: (713) 798-4649 fax: (713) 798-3729 Volume-Number: Volume 19, Number 101