[JDEV] Timestamp format (summary)

[Zoom Juice]

--- Thomas Charron <tcharron at ductape.net> wrote:
> Quoting Zoom Juice <zoomjuice at yahoo.com>:
> > This issue shouldn't be considered closed yet.
> 
>   The only reason why it's getting close to being considered closed is that the 
> majority of developers seem to want to support the ISO standard, using GMT as a 
> baseline.
>   I'd like to remind everyone, that proposals are subject to change.. ;-P

Hmm, well yes, in the interest of not getting bogged down in the minutae, I'll
make this my last post on the subject.  OK, here goes...

Setting standards by majority vote is a slippery slope.  Been there, done that.
What you wind up with is usually a platypus - an elephant designed by committee.  

The guys who put together the excellent standards on which the internet is 
based (read almost any rfc, with a few notable exceptions) didn't do it 
democratically - they did it by identifying the *best* approach, or something 
darn close to it, and pushing it though by shear force of will - by writing 
excellent docs, by laying it all out, posting the result on the net and waiting 
for the hordes of highly intelligent critics out there tear it apart.  The 
proposals that survived, sometimes in a heavily modified form, often went on 
to make a real impact. Note that this has very little to do with democracy.
It has everything to do with carrying the process all the way through - 
killing off all the ugly ducklings so that the swans survive.  Hehe, sorry
for the mashed metaphor.

Here we've all got aspirations to being involved with something that could
become another one of those core standards that that make the web what it is.
We all think that might happen, that this thing could become as ubiqitous as
a POP server, or we wouldn't spending our valuable time here.  Well, if we 
want to go all the way, this has to be *darn good*.  Not just *kinda good*.  
And that means, every aspect of it, not just parts of it.

OK, so what am I leading up to?  Well just this: invoking the letters ISO
doesn't mean something is the best it can be.  ISO has put out lots of dumb
standards that died on the vine - every standards organization has.  (For
an example of a real turkey, check out VESA's original proposal for a
protected-mode version of VBE - which required you to load the code, in
source form, in a language they dreamed up, from the disk, just to switch
video banks, modes, etc.)  It's necessary to look critically at the ISO 
standard - here are a couple of questions I have:

  - What function does the "T" perform?  Is this the *only* way we can
    find the beginning of the "time" part of the string?  Is there something
    else that could follow, e.g., "P", a "place"?  (That last question was
    rhetorical.)

  - If this is a standard, who is currently using it?  I haven't seen it
    used anywhere at all... where is it used?  Sendmail or somewhere?
    Personally, I've never seen or heard of it before.

OK, like I said, I won't post any more on this topic, I just thought it
would be worth making a few comments on the issues of standardizaion, and
the design process itself.  Timestamp format is a tiny part of the whole 
design, but every part of the thing has to be treated as if it were the
most important issue in the world.  I forget who once said that good
operating systems are good because they are "the sum of many subtle
perfections".

=====

__________________________________________________
Do You Yahoo!?
Bid and sell for free at http://auctions.yahoo.com