Re: After 5

2014-09-17 21:56, Daniel Glazman wrote:
>> In traditional programming, a “module” is a unit with well-defined
>> relations to other units. CSS “modules” are just parts or aspects of
>> CSS. They may overlap, they may be ignorant of each other, and they may
>> conflict with each other.
> <pun>
>    « CSS is not a programming language »
> </pun>

My point was that if you use programming language concepts in the 
context of other computer languages, you should use them in a manner 
that is reasonably consistent with the original concepts. XHTML 
modularization once did that and wasn’t very successful, but consistence 
of concepts was not the culprit.

The current “modularization” of CSS, as well as the ongoing 
“modularization” of HTML, is something different. It’s just splitting 
things into parts. This may well be the best you can do, but it would be 
better to call them just parts, or perhaps a base language and its 
separate extensions. This way, you would avoid confusion with modules 
proper and with all the benefits and problems that modularization has.

To developers, “base definition” and “extensions” carry a more 
realistic, more useful message, implying that extensions may be mutually 
contradictory and may contain (possibly incompatible) changes to the 
base language.

> I would like to add something about modularization we rarely think of, 
> being spec authors: a collection of lightweight specs is better for 
> Web authors than a single huge document. And I suppose implementors 
> don't really care as soon as the specs are well written. 

For true modularization, that would be true. For division into parts, or 
for base definition plus extensions, it’s rather the opposite. If there 
is a single spec, let it be 1,000 or 2,000 pages, you can check any 
single thing there and get the ultimate answer. You may need to follow 
crossreferences and spend a few bottles of your favorite cola beverage 
before you can be certain. But it’s still very different from going 
through, say, all the CSS recommendations and drafts to find an answer 
to a simple thing, possibly ending up with two or more different answers.

If there will be a base HTML definition and a registry of defined 
extensions, the situation will be somewhat better. But you would still 
have to go through all of them to find an answer to a simple question, 
unless someone prepared a useful annotated combined index of their 
content—and such an index would be a monolithic “spec” in disguise, 
without an official status, I suppose.

-- 
Yucca, http://www.cs.tut.fi/~jkorpela/

Received on Sunday, 21 September 2014 17:31:28 UTC