RE: Development standards for the World ( was RE: ASP Thresholdinformation )

> From: qsrvbas@xxxxxxxxxxxx
> 
> It seems to me that the best use of development money (where the world
at
> large is the audience, not a closed environment such as within a
corporate
> LAN where corporate standards might dictate choices) would be to spend
it
> along standard lines. If a standards body exists, then those standards
> should be followed and not the 'standard' of a particular company.

That's a nice thought, but historically speaking unattainable.  Take a
simple example: SQL syntax.  SQL has been around for about 30 years
(Oracle's product is over 25 years old), and has had three different
releases of standards (SQL89, SQL92 and SQL99).  And still they haven't
standardized the syntax for returning the first (N) rows of a query.
Because of this, each vendor has created their own extension and to
no-one's surprise, none of the vendors match.  SQL Server, Informix,
Oracle, DB2 and MySQL all have a different syntax.

With browsers, the issue is rarely in how the code is rendered.
Eventually everybody ends up agreeing on that for the most part.  This
is where Mozilla spends a lot of its time, trying to keep up with the
latest standards for things like CSS, and doing a reasonable job.

The big problem with browsers is what is called the Document Object
Model, or DOM.  Somewhere along the line, it became clear that an entire
HTML document, no matter how complex, could be represented as a set of
nodes not unlike an XML document.  By exposing this model to the
programmer via JavaScript, that allowed for some incredibly powerful
techniques.  Relatively simple code allows you to create functions that
closely mimic 5250 behavior.

The problem is that there is a dearth of standardization in this area,
and so each vendor does their own thing.  Microsoft has invested a ton
of money in this area and simply blows everyone else away, including
what I consider to be the best online documentation in the world today,
Microsoft's MSDN site.

In any case, things that can be done in IE simply cannot be done in
Mozilla today.  A simple example: I can dynamically change the user's
keystroke from a lower-case 'a' to an upper-case 'A' (just like we do in
the 5250).  This is because the character value in the keypress event is
updatable.  In Mozilla, this property is read-only.  Without this
capability, I have to choose between some pretty ugly workarounds.

Anyway, enough on this.  It's not a religious issue.  If your company
can afford you taking the time to make your code browser-agnostic, then
more power to you.  But if you don't, and you have to choose a specific
bropwser to support, then don't feel bad: it's just a business decision.

Joe