From derek@knosof.co.uk Thu May 4 17:02:04 1995 Received: from knosof.demon.co.uk by dkuug.dk with SMTP id AA18058 (5.65c8/IDA-1.4.4j for ); Thu, 4 May 1995 18:16:40 +0200 Date: Thu, 04 May 1995 17:02:04 GMT From: derek@knosof.co.uk (Derek M Jones) Reply-To: derek@knosof.co.uk Message-Id: <1143@knosof.co.uk> To: sc22wg15@dkuug.dk Subject: Re: (wg15-uk 718) (SC22WG15.506) CCTA concern about value of POSIX standards X-Mailer: PCElm 1.10 Lines: 79 X-Charset: ASCII X-Char-Esc: 29 In message <199505041314.AA13354@dkuug.dk> "Jim Isaak- isaak"@csac.ljo.dec.com, saak@ljo.dec.com writes: > > Since the POSIX project began it has received the full support of CCTA. We have > demonstrated this support by participating in WG15 and also by supporting > and participating in P1003.0. > > We currently recommend that users mandate conformance to 9945-1 when procuring > platforms for a multi tasking multi user environments. Over the past twelve > months, however, we have started to question the value of this advice to our > customers. I think the phrase "recommend that users mandate" might be part of the problem. I understand that CCTA are not in a position to "require that users mandate". Do you have any figures on how many of your users follow your recommendations? Also how do users ensure that the products they buy do in fact conform to standards? Do they consult the NISt validation report, which lists current validated OS/hardware combinations? > > The aim of the POSIX project was to promote the concept of Application > Portability at source code level. Exactly. I imagine that most of your users are removed from the source code level. So it would be unfair to claim that POSIX did not meet their needs. Interoperability it a higher level concept for which there are currently few standards > How far has the POSIX family of standards > succeeded in real applications in end- user businesses? The information at > our disposal suggests success is limited. On the one hand, departments who > have purchased POSIX compliant platforms have found the more detailed > specification of XPG4 Base necessary because the 'flavour' of the product > bought will be more consistent across a family of products. Application > portability has been achieved by the machine - specific implementations > of RDBMS and similar products allowing application portability at the > business level, but by the route of lock - in to a strategic software > product. It may not be POSIX that helps the user! > My experience in helping users and vendors test applications for conformance also suggests that XPG is closer to what they want. But this is only because X/Open can move faster in producing documents that IEEE. > Our customers have accumulated a heterogeneous array of systems and > interoperability is the key requirement. We are unaware of any way in > which the POSIX framework is progressing this. All progress seems to > emerge from the market and even X/Open finds itself under threat > from this user need. The challenge to all 'open systems' approaches is the > success of products in the market in delivering increasing business > benefit to the user. > I think that it is a big mistake to regard Posix or XPG as an all encompassing solution to your problems. The specifications contained in these documents are a means to acheiving a solution. I must also say that CCTA, along with most companies, have a very poor record of checking that the products that they buy do in fact conform to standards. The first step along the road to checking conformance to standards is to purchase copies of the appropriate documents. Are any figures available for the number of copies of the POSIX & XPG documents available within the CCTA user base? > > CCTA's view is that POSIX project is dealing with a computing environment on > the wane. The participants should take stock of the markets, of users needs > and ask themselves if they are still aiming at the right target - we would > welcome WG15's views on where they see the POSIX project going in the future. > > The view of CCTA's management team is that it is only prepared to allocate > the resources for continued participation in helping to develop a new view > to take our user policies up to the millenium. Do you have any ideas what this 'new view' might look like? derek