- From: Asynch Messaging <asynch_messaging@hotmail.com>
- Date: Thu, 20 Dec 2001 10:36:40 -0800
- To: <xml-dist-app@w3.org>
- Cc: "Rich Salz" <rsalz@zolera.com>, <Jorgen.Thelin@capeclear.com>
> ----- Original Message ----- > From: "Rich Salz" <rsalz@zolera.com> > Subject: Re: NameValue and NameValueList data types > I do the Apache name/value just fine in Python. > I don't see why this is a more compelling, more "primitive" type than, > say, most other types in the C++ STL, for example. Is it just because > CORBA had a fondness for it? I see name/value pairs as more fundamental for building and exchanging semantic information - which is a lot more interoperable than exchanging syntactical information. All you've got are names and indexes, anyway. There are more syntaxes than there are semantics - and more are built each day, while the semantics stay the same. And yes, this implies that people will be building bridges - as long as there are programmers, there will be a need for bridges. I personally don't worry too much about making the One True Syntax for something like a Hashmap - documenting the three or four most common ways of doing a Hashmap is useful to me however.
Received on Thursday, 20 December 2001 13:37:13 UTC