Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!samsung!uunet!ogicse!littlei!max.intel.com!dar From: dar@max.intel.com (dar) Newsgroups: comp.unix.sysv386 Subject: Re: Unix Support or lack thereof (long) Keywords: sco unix interactive wars Message-ID: <1629@gandalf.littlei.UUCP> Date: 4 Jan 91 17:41:53 GMT References: <5553@rsiatl.Dixie.Com> <1990Dec30.193929.16181@kithrup.COM> Sender: news@littlei.UUCP Lines: 100 [Deleted commentary on how ISC & SCO UNIX support stinks relative to support from WordPerfect for their DOS package.] I just can't let this one go by ... The whole point is that UNIX is phenomenally more complex *in its potential uses* than even a complex word processing system like WP. The cost of support is not in development [even so, it's misleading to state that "90% of UNIX comes complete from AT&T ... the sheer maintenance, version control, repackaging, publishing and so on for over 60MB of source is difficult enough even if you don't change one line. To change 10% makes it a very major deal]. The cost of support is driven up nearly exponentially (actually, the factorial) as the number of degrees of freedom in usage and capability goes up. Running a single program in a hermetically sealed environment, like WP on DOS, is a fairly straightforward support task. UNIX has hundreds of utilities, many of which are as complex, or more so, than WP. Supporting UNIX means supporting each of those utilities in possible permutations with multiple simultaneous users. Just being a system administrator in UNIX is a major, major support issue. Picking a printer file from WP just doesn't compare. In another issue, many of the UNIX utilities are very open-ended in their potential uses. For example, one can use the C compilation system to create arbitrarily complex programs for any possible use of the machine. Supporting UNIX usually means supporting people's interaction with their systems as they learn C. Compared to C, which deals with the creation of rapidly more complex logical structures, the subject matter of WP, mere text, is considerably more static. The actual experience of any UNIX software company selling and supporting UNIX is that 90+% of the support calls are "education" calls, where the problem is resolved by teaching someone UNIX. Calls generated by true bugs are extremely rare. It is extremely common for people to place support calls not to resolve an ambiguity in the manuals or operation of C compiler switches and the like, but rather to get an over-the-phone tutorial on how to use C, what a 'cast' is, how pointers and indirection work, and so on. For all of its good support, WP doesn't commit to teach people how to touch-type over the telephone, nor does it spend time teaching people how to use their word processor to correctly write a legal document or to tap out the next Great American Novel I know there are many expert users of UNIX who never use support calls to get educated, or to call a "free" support line in lieu of taking five minutes to crack open a manual. My guess is that 10% of the users are using 90% of the support traffic for self-education or to get information which is in the manual set. Somebody pays the cost of that support, and the cost (obviously) is born by the 90% of the customer base which does not utilize support for such uses. Now, I know that if you want to be in the business you got to pay the freight. Part of the cost of selling a complicated product like UNIX is to be ready to shoulder a larger support burden, probably involving a lot of customer education. The only question is how the costs of that service get allocated and who gets to pay. Ultimately, the customer base is going to pay (after all, that's where the money comes from, not thin air). So the real question is how you are going to allocate support costs through your customer base. Personally, as one of the 90% of the customer base that almost never calls support, I'd rather get a price break on my UNIX package than have to pay one thin dime more to subsidize a support service for those people who call to complain that they didn't get an "A" and "B" compiler along with their C compiler [true story]. After all, knowledge and capability are competitive advantages in this business. Why should I spend a few hundred bucks extra per UNIX package so that SCO or ISC can provide free support to those of my competitors who are too stupid or too lazy to crack open their UNIX manuals? The problem with free support is the same problem as free anything, known as "the tragedy of the commons" in game theory - free stuff tends to get abused by the wrong people. Free information calls mean people just don't bother to use phone books. Free access to "the commons" (a grazing area in the center of the village open to all, for those of you not from New England) means that the commons gets over-grazed and ruined by those too lazy to herd their sheep out of town. Free support in the style to which some would like to be accustomed simply means that an abusive minority will chew up the resource to get educated over the phone. A better solution is that people pay some fair rate for support. That way, the people who don't need support don't end up subsidizing those who do, and those who want a lot of support can pay for what they use. I know this will annoy a lot of socialists out there, but paying a fair price for what you use tends to work out very well in the real world. Finally, let me note that the $300 to $600 cost of modern "big time" DOS packages like WP is extremely high relative to what you get by way of capabilities and support. Considering that most UNIX systems deliver hundreds of more programs and a far richer set of capabilities, the fact that they cost $1000 to $2000. Sure they could be better, but SCO and ISC should be commended for the support they do provide. Dimitri Rotow PS - Sorry to lob a big note like this on the net and then move, but I will be moving this next week back down to California and will be off the net for a while. I'll try to catch up on the flames as able.