Archived from groups: alt.comp.hardware,alt.comp.hardware.homebuilt,alt.comp.hardware.pc-homebuilt,aus.computers,comp.hardware (
More info?)
On Sun, 03 Jul 2005 01:41:18 -0500, David Maynard
<nospam@private.net> wrote:
>> There is no good reason NOT to support it to the extent that
>> it works. There is no good reason not to support ANY
>> protocol, even one completely home-grown, to the extend that
>> it installs in same manner and is supported though whatever
>> features it, itself, supports in a standardized manner.
>
>That is simply silly. 'Support', of any kind, costs time and effort and
>when the return doesn't justify it then one doesn't do it.
Nope, "Support" was already there in Win9x and 2K. It
requires more time to remove it than leave it alone, so far
as functionality is concerned. If they want to drop active
technical support policies for implementing it, that is
another story and understandable.
>
>
>> Perhaps it's only antiquated in your and MS's minds.
>
>It's antiquated by comparing it's capabilities to the alternatives.
Again, no.
Adding features that someone doesn't need, does not in
itself antiquate something that is better for the user's
needs. Again it seems to come back to the idea that what
you want, specifically, should suddenly make what someone
else wants, irrelevant.
>
>> Certainly not so to anyone who uses it.
>
>Neither are buggy whips, to those who 'use them'. but there's little reason
>to 'support' them in "horseless carriages."
So apparently you are choosing the buggy whip. It has more
versatility than a car, since it whips animals and
unfortunately slaves (at the time). You choose this instead
of the faster alternative.
When one does not need the TCP/IP features, it's simply
counter-productive and illogical to use it when those very
features make it slower.
>
>> That doesn't even
>> begin to mean someone else shouldn't use something that
>> suits their needs more than your own personal alternative
>> would.
>
>They're free to use it. They'll just have to provide their own 'support',
>or a third party who might be interested.
Yep, but that has nothing to do with "antiquated". What
will antiquate Netbeui is when an alternative exceeds it's
performance but also is non-routable. In other words,
TCP/IP can never possible be the thing that antiquates it,
because both exist for different reasons.
>
>> The fact of the matter is that the main benefit of TCP/IP is
>> that it's routable..
>
>Ain't no 'little' thing.
Right- BUT if you don't need to route, perhaps even don't
WANT to route, then it is not only a feature that one
wouldn't miss, they choose to avoid it and reap the
performance gain from doing without.
>
>Not to mention it's universally supported across virtually every platform
>on the planet instead of a single O.S..
... which is completely irrelevant in fixed scenarios where
no routing is needed. Nobody is arguing to abandon TCP/IP
for uses where it is needed, only to use the right tool for
the job.
>
>> Once you remove that need, you're left
>> with a slower protocol.
>
>Remove it and you've got an antiquated, severely limited, protocol.
It is senseless to suggest use of a protocol with more
features when none of those features are needed and it's
slower because of them.