Re: Just say no to threads [Was: Software architecture]

From: Nick Landsberg (SPAMhukolauTRAP_at_SPAMworldnetTRAP.att.net)
Date: 11/02/04 

Date: Tue, 02 Nov 2004 01:30:03 GMT

David Lightstone wrote:

> "Nick Landsberg" <SPAMhukolauTRAP@SPAMworldnetTRAP.att.net> wrote in message
> news:mzvhd.820149$Gx4.582235@bgtnsc04-news.ops.worldnet.att.net...
>

[SNIP]

>>>
>>
>>Yes David, they still do exist, and the larger the organization the
>>more likely it is to occur. As one example from recent
>>memory, the hardware types specified that the next release
>>of a board would be diskless and that there would
>>be two boards in the shelf which had disks from which
>>the other 10 boards would boot over NFS. This was
>>thrown over the wall the to software types who had to
>>implement it. The software types eventually pointed out
>>that NFS could not meet the 5-9's reliability requirements
>>even with two disks on the shelf.
>
>
> Somebody tried to implement before the design was analysed for consistency
> with the requirements (5-9 in this case). Definitely not a waterfall type
> development (there would have been an analysis and a review).

Very perceptive, David. Unfortunately, in that particular
corner of the corporation, outside review was "optional."
The outside review team was called in by the software types
*after* they came to the conclusion that the hardware types
had sold them down the river. (Note that "outside" in this
case refers to a team of subject matter experts from inside
the corporation but "outside" the organization(s) involved.
It's an awfully large corporation, so that you can usually
find folks to fit that bill.)

>
>
> The application wouldn't perchance be a cell phone base station now would it
>
>
You have proven in the past to be a very perceptive
person, David! :) Otherwise, "no comment."

[ More SNIPPED ]

Now, to try to bring this discussion back to the subject(s)
of the newsgroups. ...

OPINION: There is more to "software engineering," and "objects"
and programming in general (whether XP or not) than providing
the basic functionality. Any methodology which ignores (or sublimates)
the non-functional (hidden) aspects (like performance, reliability,
maintainability, etc.) is missing the boat in being able
to provide large-scale systems. Any methodology which does not
have pretty good feedback loops between the folks who talk
to the customer(s) and the developers, is also missing the
boat. When the folks on one side of the wall and the other
side of the wall have differences, most often there is
"friction" (if not outright warfare). Any methodology/process
which fosters communication between the developers and the
folks who have the final say on the (hidden) requirements is
beneficial, in my opionion. But it should *not* be considered
an "end all and be all."

NPL

Relevant Pages

  • Re: folk guitar workshops on archive.org
    ... innocuous, but what happened to folks being asked to participate here, ... David Enke, etc., that actually abide by the charter? ... playing Folk Guitar, which I actually found quite interesting. ... Patrick in this discussion. ...
    (rec.music.makers.guitar.acoustic)
  • David Voisine or whatever, spam king
    ... But you see folks it's in this guys nature to do such ... So I change my statement to I hope you david, ... "Their pictures match their crappy attitude." ... or have kids in the room.. ...
    (sci.med)
  • Re: Boot Camp Update March for Friday 23rd
    ... On Mar 24, 9:35 pm, David J Dachtera ... I'm sure some folks would be glad to purchase a machine out of ... Unofficial OpenVMS Hobbyist Support Page:http://www.djesys.com/vms/support/ ...
    (comp.os.vms)
  • Re: RubyConf in Second Life?
    ... I think you're probably right about this David. ... participants interacting with SL participants - it would become much ... with effectively infinite bandwidth like a University). ... > I'm still wondering what exactly the SL folks will be getting out of the ...
    (comp.lang.ruby)