W3C home > Mailing lists > Public > www-dom@w3.org > January to March 2001

Re: conformant function names

From: Joseph Kesselman <keshlam@us.ibm.com>
Date: Tue, 30 Jan 2001 11:32:13 -0500
To: www-dom@w3.org
Message-ID: <OF5C3C0F86.DED6CCEA-ON052569E4.005AB39E@pok.ibm.com>
I accidentally answered Dieter's question on the IG list rather than on the
public list, where he asked it. Reposting:

>Is an implementation in other languages than Java or ECMAScript still
conformant
>if totally different names are chosen?

By our current definition, I think the answer is "yes, but this would fail
my own criteria for quality-of-implementation." Part of the goal of the DOM
is to minimize relearning as people move between implementations. That
implies that renaming should be absolutely minimized, and that idiomatic
compatabilty with the DOM spec is more important than idiomatic
comptability with other classes in the same system.

If there's a keyword conflict, I would recommend a minimal and systematic
change -- which means finding all the conflicts, then considering whether
there's a simple and obvious transformation that can be applied to those
names to resolve them.

If you really want both count() and length(), write both into your
implementation... but length() will be preferred by anyone who may want to
reapply their DOM code elsewhere.

______________________________________
Joe Kesselman  / IBM Research
Received on Tuesday, 30 January 2001 11:32:36 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 22 June 2012 06:13:48 GMT