Path: utzoo!attcan!uunet!snorkelwacker!usc!wuarchive!sdd.hp.com!hp-sdd!hpcndnm!jad From: jad@hpcndnm.cnd.hp.com (John Dilley) Newsgroups: comp.sys.hp Subject: Re: Software patches Message-ID: Date: 10 Oct 90 16:44:54 GMT References: <18@gauss.mmlai.UUCP> <5570507@hpfcdc.HP.COM> <24818@uflorida.cis.ufl.EDU> Sender: news@sdd.hp.com (Usenet News) Organization: Hewlett Packard, Colorado Networks Division Lines: 77 In-Reply-To: bb@palmetto.cis.ufl.edu's message of 10 Oct 90 03:03:13 GMT Nntp-Posting-Host: hpcndnm.cnd.hp.com In article <24818@uflorida.cis.ufl.EDU> bb@palmetto.cis.ufl.edu (Brian Bartholomew) writes: >In article <5570507@hpfcdc.HP.COM> rer@hpfcdc.HP.COM (Rob Robason) writes: >>2) What happens if a customer chooses a standard procedure of >> installing every patch that HP makes available? Patches are really >> intended, and QA'ed, for installation on a "standard" released >> system. There may be unexpected interaction between two or more >> patches. We would like to limit patch distribution to those who >> must have them to reduce the risk of introducing unexpected >> interactions between these patches. The last thing we want to do is >> have a critical problem introduced at a customer installation >> because of a patch; for one thing it is immensely more difficult to >> isolate on a non-standard (i.e. patched) system. I think this is >> at least one reason for our reluctance to give these patches to >> everyone. >This is the most damning admission I have heard on this group in >several months. You have admitted that HP can not even keep their own >bug fixes straight in their own house. Further, if HP is not >confident enough in their own repairs to known they will work with >their own software, then they aren't really repairs, now are they? I think Rob's point is that our patches are QA'd on the previous release, not on top of the previous release with all available patches on it. To truly guarantee that a patch will not break a customer system, HP would have to test the patch on all of the following configurations: - the last "pure" QA'd release upon which the patch is supported; - the last release with all current patches applied; - the last release with every non-empty subset of patches applied. (Yes, I know, the second case is subsumed by the third :-). The combinatorics of this get very ugly (2^n) if many patches (n) go out. We keep the patches straight internally, but on limited configurations. >Perhaps in the HP-UX technical support sales literature, HP should >mention that the customer may have his choice of known bugs; but by >company policy, in no case shall he be allowed to have a system >completely free of bugs. Now that I read that again, perhaps it should >be moved to the Sales Manual for the technical support managers... HP does make every (reasonable) attempt to ship software free of bugs. When serious bugs do make it out in software (nobody is perfect) there is a need for a patch process to fix them. It's either that or hold up the release until all the bugs have been found and eliminated. In my understanding, though, only very serious bugs will get patches distributed. And patches go through our field force so the field can know what's going on, and can do their jobs supporting them. Given your above statement, however, I think you may fall into a set of customers whose needs HP may not understand well enough. Do you do your own support, or do you rely on HP support? In some environments having the "latest and greatest" is more important than having stable, known configurations (I used to be a university CS department admin, I understand the needs of that type of environment). For many of our major customers, however, a stable, known configuration is required. I don't think we can necessarily solve the needs of both environments with one solution. I also don't know if "unsupported patches" are something HP would consider ... but perhaps that's a potential solution? Personally speaking, I would love to see HP be one of the most responsive hardware and software suppliers while keeping up the level of HP quality. Keep the ideas coming ... someone is certainly listening... -- jad -- John Dilley Hewlett-Packard Colorado Networks Division UX-mail: jad@cnd.hp.com Phone: (303) 229-2787 -- This is not an official statement of Hewlett-Packard Corp, and does not necessarily reflect the views of HP. The information above is provided completely without warranty of any kind.