khb%chiba@Sun.COM (Keith Bierman - SPD Advanced Languages) (10/20/89)
>psmith@convex.com (Presley Smith).... >....This vote is good news for the ENTIRE FORTRAN community. ... Nice hyperbole. I take it that you do not count the 2/3 of X3J3 that voted the draft out of committee as not being members of the FORTRAN community ? >FORTRAN 77 will continue to be an active ANSI Fortran Project so that >interpretations may continue to be made on that standard by X3J3, but >no major updates to that standard are anticipated. ^^------------------------------------------------ Not anticipated ? There is nothing to prevent it. And, "interpretations" of X3.9-1978 which conflict with the ISO standard Fortran are (in the long run) quite likely. >The X3 committee voted today by a large margin to retain ANSI X3.9-1978, As I understood things, ISO tasked ANSI to write the next rev of the world standard for FORTRAN (chose your capitialization to taste) way back '78ish to follow up the work of the last committee ('77). X3 tasked J3 to do this (i.e. "own" the future of FORTRAN). So what X3 has effectively done is to repudiate its subcommittee. Of course, X3, tasked to preside over the standards _process_ (not actual standards), has a much better understanding of the FORTRAN communities needs than J3. >Internationally, the Fortran Extended standard will be the SAME as the Unless WG5 as agent for ISO changes its tune there will be no Fortran Extended Int'l standard. There will simply be an international Fortran standard. I am not foolish enough to think this is in NO ONEs best interest; clearly there are those who feel quite strongly that X3.9-1978 is the "right thing". But what are the consequences for the rest of us ? 1) There will be 2 US standards. If for no other reason than clever vendor reps will finagle a FIPS for it. Thus there will be 2 validation suites. Given the "play" in Fortran standards documents it is quite unlikely that it will be possible to pass both validation suites with the same code (of course compiler flags "solve" this problem). RESULT: Higher cost of developing compilers. Will retard development of ISO compliant compilers by US vendors strapped for resources. Probably result in lower quality of implementation for _both_ standards. 2) As pointed out by others, the International community is unlikely to view the X3 act as responsible. But we americans have lived down worse embarassments ... of course one might suspect that ISO might not be willing to entrust future work to X3 (I wouldn't, if I were in their shoes). RESULT: More foreign travel for folks involved in standards work. RESULT: There would appear to be little reason for WG5 to worry about X3J3 activities anymore; this might enable quicker International adoption now that the US is out of the picture. RESULT: Library vendors will have a strong disincentive to provide features which rely on ISO-compliant features ... no matter how much better it would be supporting two source trees per machine is costly. Of course, this means reduced functionality for users. I am fairly certain that ISO will ordain a single version of Fortran (at least one hopes that some bodies operate rationally); and that some US vendors simply will not have the resources to do both American standards. Since it is safer to bet on Fed. Gov't mandating the obsolete standard, those vendors will go with X3.9-1978.... of course this will severely impact their ability to compete overseas ... thereby helping increase the trade deficit (but we can fix that with trade barriers, eh ?) Who will benefit ? 1) Users whose codes were written on a single system and never ported them; they can continue to use their systems with faith that their environment will never improve (and I've worked at sites where that was clearly a design goal ... and one can always argue that fixing up the environement only robbed resources from solving the real problems the organization was tasked to solve... and for any given Lab this is likely to be true). 2) Vendors who have dominated their market segment with their x3.9-1978 extended compilers. Adopting a single (new) standard would have promoted competition in each niche. I can't come up with a third group. But then it is self-evident that the union of these two groups constitutes the ENTIRE FORTRAN COMMUNITY. A wonderful day indeed. But then up here in the Bay Area we've learned to really enjoy "Turf Surfing" Keith H. Bierman |*My thoughts are my own. !! kbierman@sun.com It's Not My Fault | MTS --Only my work belongs to Sun* I Voted for Bill & | Advanced Languages/Floating Point Group Opus | "When the going gets Weird .. the Weird turn PRO" "There is NO defense against the attack of the KILLER MICROS!" Eugene Brooks
bobal@microsoft.UUCP (Bob Allison) (10/24/89)
In article <126584@sun.Eng.Sun.COM> khb%chiba@Sun.COM (Keith Bierman - SPD Advanced Languages) writes: > > [...] And, >"interpretations" of X3.9-1978 which conflict with the ISO standard >Fortran are (in the long run) quite likely. > Gosh, I hope not. If so, then 8x is clearly not upward-compatible with 77, which is what everyone is claiming is a high priority. >>The X3 committee voted today by a large margin to retain ANSI X3.9-1978, > >As I understood things, ISO tasked ANSI to write the next rev of the >world standard for FORTRAN (chose your capitialization to taste) way >back '78ish to follow up the work of the last committee ('77). X3 >tasked J3 to do this (i.e. "own" the future of FORTRAN). > >So what X3 has effectively done is to repudiate its subcommittee. Of >course, X3, tasked to preside over the standards _process_ (not actual >standards), has a much better understanding of the FORTRAN communities >needs than J3. > Well, this is something of a problem for X3J3: X3 has lost all faith with X3J3. This is partially due to the fact that we're in the twelfth year of a five year project proposal (and they had to specially interpret the rules to get around the ten year rule), and partially due to the obviously divided nature of the committee. Remember, X3J3 was relatively close to coming out with a standard around 1982, when they changed direction substantially (thus the demise of "core plus modules"). I think X3 views X3J3 as just too unpredictable and essentially unreliable. And, it might be argued that X3 does have a fairly good grasp of the Fortran community and its needs (well, at least as good as X3J3's). > >2) As pointed out by others, the International community is unlikely > to view the X3 act as responsible. But we americans have lived > down worse embarassments ... of course one might suspect that ISO > might not be willing to entrust future work to X3 (I wouldn't, if > I were in their shoes). > >RESULT: More foreign travel for folks involved in standards work. > It's a rough life, but someone has to do it ;-) Seriously, this is significant, if true. Already, it is nearly impossible for an individual to attend meetings without outside funding. Although this premise does seem somewhat extremist. ANSI foots the bill for a lot of ISO standards work, and I'm not sure where ISO is going to come up with the cash to administer all these language standards they're going to take away from ANSI. Also, despite all the disagreements, the US is a significant factor in retail compiler sales. I don't think the ISO would work to encourage a rift between the US and ISO (no matter how powerful you believe the EEC will be, they will still want to sell into the US market and vice versa). >RESULT: There would appear to be little reason for WG5 to worry about > X3J3 activities anymore; this might enable quicker > International adoption now that the US is out of the picture. > Well, they want an ISO standard which ANSI will approve, so I don't think they'll take the current draft and go crazy. And they might still allow X3J3 to finish tweaking it based on public comment, so as to allow for ANSI approval of the standard (otherwise, ANSI will have no choice but to reject the draft). >RESULT: Library vendors will have a strong disincentive to provide > features which rely on ISO-compliant features ... no matter > how much better it would be supporting two source trees per > machine is costly. > This will be true anyway as long as there are still platforms without 8x. I'm not sure whether this action made a significant difference in the time line. Suffice it to say that I honestly believe this is a tempest in a teapot. I am glad that 77 will continue to be a standard, but I don't believe it will affect the adoption of 8x (by coders) significantly enough to tilt the scale one way or the other. It will some delay vendors from bringing out a product until they are sure of a market, but that just allows aggressive, risk-taking companies to try to cash in on the potential rewards of being there first. For instance, I can say with assurance that this hasn't changed our development plans for 8x at Microsoft one iota. Bob Allison