home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.unix.sysv386
- Path: sparky!uunet!usc!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!uchinews!machine!chinet!ignatz
- From: ignatz@chinet.chi.il.us (Dave Ihnat)
- Subject: Re: DELL SVR4 - separate /usr and /var filesystem?
- Message-ID: <BuKsLM.IxI@chinet.chi.il.us>
- Summary: Right, thanks for supporting the specious unbundling argument...
- Keywords: dell svr4 filesystem bundling
- Sender: Dave Ihnat
- Organization: Chinet - Public Access UNIX
- References: <id.I13T.6_D@ferranti.com> <BuEvwy.2rt@gator.rn.com> <2AB40B3A.4AC7@telly.on.ca>
- Date: Mon, 14 Sep 1992 16:03:21 GMT
- Lines: 96
-
- In article <2AB40B3A.4AC7@telly.on.ca> evan@telly.on.ca (Evan Leibovitch) writes:
- >The real choice is whether to pay for it or not. If you know ahead of
- >time you'll never need it, being forced to buy and then de-install it
- >is not what I'd call a superior "choice".
-
- Ah; you must be a young pup, and have bought the argument that IBM and others
- proposed when they first started unbundling software. You obviously haven't
- been in enough real-world sites to have had to curse the day they unbundled
- everything, so that when a need arose you had to suddenly justify spending
- hundreds or thousands of dollars to get the tool that, had it still been
- bundled as part of the package, could have solved the problem in minutes.
-
- This isn't just babble--look at Unix. Used to be that, no matter how tight
- the budget, if they were running Unix you knew that you could turn out good
- quality printed documents, since all versions had n/troff. Today, you either
- talk them into spending money on a document preparation system--not likely
- at most sites--or find a DOS system with some word processor. An inferior
- and degraded state, IMN-S-HO.
-
- >The need for some people to run DOS programs under UNIX is still very
- >real, and is not likely to go away for some time. But technically
- >superior options, such as a separate MS-DOS server running Desqview/X,
- >and an increasing number of DOS software being ported to run native
- >under UNIX, make this option less necessary than it used to be.
-
- "Technically superior options" often aren't options. There are situations
- in which the need is real, but the money needed to implement the "superior"
- solution is needed for tasks higher on the priority list. The presence, for
- instance, of Merge would allow an interim solution immediately--and for no
- (and I mean NO) additional or incremental cost. Without this option, the
- answer is usually "wait."
-
- >Yes, I understand the economics of it. Dell (or any vendor) can get a
- >significant break from Locus if Merge is bundled with every copy of its
- >UNIX. That makes the price of the product lower than it might be for
- >those who want everything, but guarantees the total product will be more
- >expensive than it need be for those who *don't* want Merge.
-
- Oh, balderdash. Then how is Dell cheaper *with all of this software
- bundled into it* than, say, SCO? If you then go on to argue that it could
- be even cheaper if they didn't do it, I'll tell you that just aren't the
- right customer for the Dell product. Which I am doing. The Dell product
- offers a remarkable bargain for the installation which needs both a reliable
- kernel, and the flexibility and features to provide an extensible production
- *and* development environment without further cash expenditure. It's a
- big win for system integrators and consultants, especially. If you have a
- restricted environment, know exactly what tools you need and will never have
- a request to grow beyond what you get with the basic system, and are assured
- that there will be no raised eyebrows when you tell your client/employer
- that they have to spend between $500-$1500 or more to buy the unbundled
- package to do what they later ask of your installation, then by all means
- buy an unbundled system. Just don't try to sink *my* raft by bad-mouthing
- the bundled solution...
-
- >Just as I don't want MS-DOS bundled into the hardware I buy (thus
- >increasing its cost) if I never plan to use it, I don't DOS bundled into
- >my UNIX. Sure that's what Locus would like -- guess why? -- but this is
- >a clear case of where unbundling is a feature, not a bug.
-
- Without going into detail, suffice to say that (from what I've been told)
- Dell doesn't really care if it feathers Locus' nest or not these days. Again
- I say, it obviously doesn't have any significant effect on the cost of the
- system to you if the bundled cost is so exceedingly competetive with the
- unbundled systems. It doesn't get in your way if you don't use it; if the
- disk space (minor, at that) required for it to quietly sit there and do
- nothing bothers you, a simple session with the system admin shell to uninstall
- it will remove even that irritant. I'll bet you're one of those people who
- remove a lot of the Unix commands in /bin and /usr/bin because *you* never
- use them...
-
- >I can understand arguments of how LAN software and the development
- >system should not be unbundled, especially when they incrementally add
- >so little to the cost of the base product. But DOS emulation is clearly a
- >third-party adjunct that is not a part of the base package, and has no
- >business being wrapped into it.
-
- Oh, my back! Because *you* need networking, you're willing to accept the
- possibility of incremental cost of bundling *that* capability, but not of a
- feature that seems useless to *you*?
-
- If I seem more vehement than is appropriate, it's because I'm seeing not only
- the same stupidity that justified unbundling of Unix in the first place, and
- IBM systems originally; but I'm seeing somebody arguing from a narrow point of
- view, in a public forum, for a position that could be used by vendors to ruin
- products that *I* have an interest in seeing continued.
-
- Between OS/2 and the (promised) features of NT, if we don't find a way to
- provide a sophisticated, capable, and cost-effective Unix that clearly offers
- something more for its cost than these single-user alternatives, then we'll be
- in the same boat we found ourselves when the PC was introduced in 1981--
- stepping back to technically inferior products for a number of years because
- of marketing and dollar decisions, and not for technical reasons.
- --
- Dave Ihnat
- ignatz@homebru.chi.il.us (preferred return address)
- ignatz@chinet.chi.il.us
-