Re: atx

Sean B. Palmer wrote:
> * Why not use an HTML WYSIWYG editor?

1. There aren't any.
2. People are very attached to their text editors.
3. Text editors are far more common than WYSIWYG editors.
4. WYSIWYG editors are less natural than atx.
5. WYSIWYG editors can't be used with other Unix tools.

atx solves all of these problems, and I don't think it introduces any 
new ones. (Can you think of any it does?)

> Heh, but did you see the XHTML 2 Considered Harmful thread?

Yeah, lots of blog coverage.

> I guess that I could embed UTF-8 encoded unicode characters...

Yes, you really should upgrade to a UTF-8 compatible system.

> So the design manifesto for atx contains something like "ignore one of
> the few good bits of good semantic design in HTML just because it's
> not supported by browsers yet"?

Heh!

> You could always have the engine convert the blockquote citation 
> format into something that browsers *can* pick up on, like a hyperlink 
> just under the blockquote.

Alright, alright. To be honest I was considering it. How about:

    If you're a new implementor, you'll be shocked at how badly [822, 
the email
    format] was designed. Extracting even the simplest information from a
    message--the author's address, for example, or the sending date--is
    excruciatingly painful. And I see no sign that we'll ever be rid of 
the
    horrors of 822 syntax; how can we convince users to convert their old
    mailboxes to a sensible new format?
(from Internet mail message header format: http://cr.yp.to/immhf.html)

as a format?

>>> And text/plain isn't meant to have formatting characters
>>> spewed through it: that's part of the defintion of text/plain.
>> Who said it was text/plain?
> "Content-Type: text/plain; charset=utf-8"
> --HEAD http://www.aaronsw.com/2002/atx/intro.atx

Hm. It should be something like text/plain; format=atx

>>> I want to use features such as <sec> in XHTML 2.0, and
>>> it's difficult to do that in text
>> Can't it be implied with blank #+ lines?
> Not sure what you mean by blank #+ lines.

Header lines with no header.

> Just in case we're talking past each other, I meant something like 
> this:-

Oh, I see. Hm. I guess we could do indentation.

>> On what basis is [atx] inferior?
> Well, for a start, if you're generally serving it as HTML, you need to
> keep the atx source up to date with the generated HTML version. So the
> steps for creating a document are: write atx, save atx, convert
> atx--and I know from experience that I can't be bothered to convert
> source and constantly maintain two files for one document. It always
> feels to be like a waste of drive space to have two such similar files
> next to one another...

It's easy to have Apache do it for you, or just generate it on the fly 
if there's not much load.

>>> you might as well propose a strict subset of HTML 4.01.
>> Why?
> the point is that *HTML* is very widely deployed

HTML _readers_ are widely deployed.

> it's possible to create a nice subset of HTML 4.01 with stuff like 
> SHORTTAG YES

I'm sorry, but I'm not really interested in writing another <p> or 
</blockquote>. I'll have my computer do it for me.

>> = no more XML! (bwahaha)
> Well, HTML 4.01 isn't XML either :-)

Close enough for RSI.

> you might as well make it really decent

Can you be more specific by what you mean by decent? I am quite curious.

> you might as well make it semantically rich, but that'd require 
> deploying a rendering engine,

Well, like XSL-FOs, you can render to a non-semantic language to avoid 
redeploying.

> It's such a shame that XML sucks, because the underlying 
> power-to-the-people principle is otherwise pretty good.

XML has a power-to-the-people principle?
-- 
Aaron Swartz [http://www.aaronsw.com/]

Received on Tuesday, 4 February 2003 14:38:07 UTC