chris@mimsy.umd.edu (Chris Torek) (08/25/90)
In article <26196@bellcore.bellcore.com> mo@messy.bellcore.com (Michael O'Dell) writes: >There was also a real-time clock chip which was used primarily >to remember the time when the machine was turned off. (Is seems >obvious that a million-dollar computer would be expected to provide >at least the functionality of a $10 digital watch, but people >took some serious convincing of that...) I am constantly amazed at the machines that do *not* have these things. If I have to pay more than a few thousand dollars for computer equipment, I really am not going to mind paying a hundred more for something that can tell (at least approximately) what time it is. $100 should certainly cover the cost of designing in some random watch chip + glue logic + battery + diode-to-keep-battery-from-overcharging.... -- In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 405 2750) Domain: chris@cs.umd.edu Path: uunet!mimsy!chris
aglew@dwarfs.crhc.uiuc.edu (Andy Glew) (08/26/90)
> (Is seems >>obvious that a million-dollar computer would be expected to provide >>at least the functionality of a $10 digital watch, but people >>took some serious convincing of that...) > >I am constantly amazed at the machines that do *not* have these things. At a XYZZY User's Group meeting question and answer session: A customer asks the XYZZY representative why XYZZY don't have battery backed up real time clocks. The XYZZY representative says "We don't feel that our customers need this functionality". The customer asks the audience of XYZZY customers "Who wants a real clock?" 99% say they do. Next Monday design of real-time clock module begins. Part of the trouble is that market research in the computer market isn't all that sophisticated. The market research questionnaires from the vendor are usually filled out by the purchasing office, or some executive. All that management knows is that engineering doesn't like the XYZZY machine, but they don't necessarily know the exact reasons. And minor problems like not having a real-time clock tend to get lost or forgotten in the list of complaints about the FORTRAN compiler, OS stability, etc. Sometimes you have to rank the desirability of features not only by their customer work-stopping importance, but also by their annoyance factor, and the ease of fixing the problem. -- Andy Glew, a-glew@uiuc.edu [get ph nameserver from uxc.cso.uiuc.edu:net/qi]
colin@array.UUCP (Colin Plumb) (08/28/90)
>>> (Is seems >>> obvious that a million-dollar computer would be expected to provide >>> at least the functionality of a $10 digital watch, but people >>> took some serious convincing of that...) >> I am constantly amazed at the machines that do *not* have these things. When I was at Cogent Research, we threw on a time-of-day clock and 32K of battery-backed-up RAM mostly because all the PCs and Macs we used had them and if we're going to charge $xx,000 for our box, we ought to have the toy features, as well. The audio circuity was there on an even thinner excuse. We spent some NRE time isolating noise from the audio circuits, choosing the speaker and designing the case to accomodate it, (it's a 3.5" speaker, facing straight out the front of the machine, and mounted on an internal panel so the bass doesn't "leak around" and get lost... it sounds amazingly good after listening to a Mac), and I think they'd do it again. All these features have debugging value (diagnostic routines and fatal error handlers log their activity in NVRAM for post-mortems), and they're easy to add: just schlung 'em on the same 8-bit bus you're using for serial ports, floppy controllers, and the like. Maybe one day they'll use part of that NVRAM for a write-behind buffer for the disks. Mostly, it's handy to have around, and the cost doesn't compare to 15ns SRAMS and transmission-line analysis and all that; it's a handful of commodity parts and easy, 1-MHz engineering. -- -Colin