Actually, the avowed purpose of most standards organizations is to select
specific practices in order to promote interoperability. Often this is
selecting a pre-existing practice, sometimes it is defining new practice
(which usually doesn't work as well unless it's an extension of existing
practice--cf. ANSI X3.64 and the like, which were helped by being extensions
of the DEC VT100's command set).
> Certainly one of the normal working rules
> is to make every reasonable effort to avoid colliding with existing
> practice.
True; however, if a standards group issues a draft for experimentation and
discussion, with specific language which prohibits people from treating it
like a standard (i.e., *any* Internet Draft such as the HTML 3.0 draft), it
does not then seem entirely incumbent upon such a group to conform to
practice which results from the implementation of that document as if it were
a standard.
Don't get me wrong; Netscape's implementation of tables is very nice, and
serves a real need. However, (and this is an honest question, not simply a
rhetorical one), does this mean that other companies can similarly expect to
have their experimental implementations of proposed features incorporated
into the standard? This would certainly make some of our engineers happy :),
but I'm not sure it would be best for the state of HTML as a whole. Then
again, maybe it would--the best designs tend to come out of small groups of
people who actually build things. I know that the more time I spend on this
mailing list, the more I sympathize with Netscape--I grow ever more tempted
to just implement what *I* think HTML should be, and let everyone else
follow if they decide they like it.
If existing practice is sufficient, there is no need for standardization.
The observable fact is that current practice in HTML is *not* sufficient,
as we see ever-expanding spheres of private innovation (in part because
the standardization efforts in this area are moribund--how long has HTML 2.0
sat "in committee" now, much less future revs?).
If we are to base standardization decisions on simple popularity, then let's
be frank about doing so and let the market work. This is what Netscape is
doing, and quite successfully. People are using Netscape "enhancements" all
over the net. I expect people to do the same with some of our features
(seamless Japanese support, upcoming ISO 10646 UCS-2 support, and others
I can't talk about in a public forum).
Amanda Walker
Software Architect
InterCon Systems Corporation