The future of HTML - enlighten me?

[Dustin Boyd]

XHTML 2.0 is no longer being developed and the decision has been made

to support HTML 5, which has an XML serialisation.  With HTML 4.x, XHTML

1.x and even through some of the various drafts of XHTML 2.0, I was

able to take in the information and use the various elements and

attributes.  With HTML 5, it's simply too large.  I realise that XHTML

2.0 was moving too slowly, but the simplicity made it rather nice.

HTML 5 on the other hand appears to be going the way of a programming

API, where you need a constant reference, reducing the usefulness.

Honestly, a lot of features are being asked for, but how many of those

requested features will be used enough to warrant their additions in

practice?

I know the same argument about HTML 5 being "bloated" has been made

before, and I also know that alternatives deemed as being "viable" are

rarely made.  However, it needs to be cut down somehow.  It's too

complex to be considered practical.  At this point, removing features

seems to be a ridiculous idea.  Starting from scratch isn't very

appealing either.  If features were to be removed, how would anybody

decide which features to omit?  It seems as if several elements and

attributes are intertwined with others.  This makes the prospect of

removing features seems rather implausible.

Oh, and while I like the idea of the CANVAS element, does such a thing

really belong in a markup language used for conveying information?

After all, the canvas API could be retained since it would rely on an

external implementation such as JavaScript, but why couldn't something

like the OBJECT element be used rather than a dedicated element?

Is HTML 5 actually being defined as a markup language as the name

suggests or is it turning HTML into a programming language?  If it's

becoming a programming language, why is it still defined as "HyperText

Markup Language" rather than "HyperText Programming Language"?  Is

HTML 5 really headed in the right direction?  Should it be split into

different sections to separate the programming from the markup?  Or

perhaps those who want a simple language should just use XSLT to

transform XML documents to HTML 4.x or XHTML 1.x.  After all, HTML 5

doesn't seem to be for authors.  That is not to say that HTML 5 is

completely useless.  I'm merely trying to make a point of the fact

that HTML 5 is more like "HTML for browser vendors".  The "Author"

view for the HTML 5 specs on the WHATWG site is rather useful, but it

still has a lot of things that appear to be more beneficial to

implementors than to the authors that create the documents.  Again,

this makes me wonder whether HTML 5 is being defined as its name

suggests or whether it is becoming a programming language instead of a

language for conveying information.  Is there going to be another Dark

Age for the Web?  HTML 4's strict definition made an effort to remove

the things that created the first Dark Age, things like FONT elements

and other presentational items as well as inaccessible items like

FRAMESETs.  In my opinion, HTML 5 as a whole is a regression rather

than a natural progression.

Of course, all of this is purely my opinion.  Feel free to direct me

toward answers that can perhaps enlighten me to the true nature of

HTML 5 because all I can see, even with my open mind, is something far

too complex to use.

[Aryeh Gregor-2]

On Tue, Nov 3, 2009 at 9:45 AM, Dustin Boyd <rpgfan3233@...> wrote:
> XHTML 2.0 is no longer being developed and the decision has been made
> to support HTML 5, which has an XML serialisation.  With HTML 4.x, XHTML
> 1.x and even through some of the various drafts of XHTML 2.0, I was
> able to take in the information and use the various elements and
> attributes.  With HTML 5, it's simply too large.  I realise that XHTML
> 2.0 was moving too slowly, but the simplicity made it rather nice.
> HTML 5 on the other hand appears to be going the way of a programming
> API, where you need a constant reference, reducing the usefulness.

One of the major goals of HTML5 is to support complex web
applications.  The spec was originally named Web Applications 1.0, in
fact.  It does have plenty of nice new features for regular authors
too, though, like <video>/<audio>, and better specification of
existing functionality to improve interoperability.  Also (arguably)
the fact that you can use non-XML syntax is good for authors, although
of course some people think XML is better.

> Honestly, a lot of features are being asked for, but how many of those
> requested features will be used enough to warrant their additions in
> practice?

All of them.  Otherwise they wouldn't have been added.  However, there
are a bunch that are only relevant to scripted pages, of course.

> I know the same argument about HTML 5 being "bloated" has been made
> before, and I also know that alternatives deemed as being "viable" are
> rarely made.  However, it needs to be cut down somehow.  It's too
> complex to be considered practical.

Some features have been split out into separate specifications (e.g.,
Web Storage), and more might be in the future.  The spec is definitely
too complicated for the average author to understand completely, of
course, but the large majority of authors don't have to read the spec
itself.  Also, it becomes much shorter if you use the whatwg.org
version and click "Hide UA text".

> At this point, removing features
> seems to be a ridiculous idea.  Starting from scratch isn't very
> appealing either.  If features were to be removed, how would anybody
> decide which features to omit?  It seems as if several elements and
> attributes are intertwined with others.  This makes the prospect of
> removing features seems rather implausible.

Thankfully, we have a talented, hard-working editor who's willing to
put in the work to do stuff like split things out into separate specs.

> Oh, and while I like the idea of the CANVAS element, does such a thing
> really belong in a markup language used for conveying information?

HTML5 is not only for conveying information, it's also meant to help
web apps.  Only a subset of it is useful for traditional static web
pages.  Splitting out most of <canvas> into a separate specification
is being considered, last I heard.  FWIW, I'm pretty sure <canvas> is
its own element rather than <object>-based or such because it was
originally reverse-engineered from a WebKit-specific element, and the
WebKit developers chose to call it <canvas> rather than using
<object>.

> Is there going to be another Dark
> Age for the Web?  HTML 4's strict definition made an effort to remove
> the things that created the first Dark Age, things like FONT elements
> and other presentational items as well as inaccessible items like
> FRAMESETs.  In my opinion, HTML 5 as a whole is a regression rather
> than a natural progression.

HTML5 goes further than HTML4, and makes <font> and <frameset>
completely non-conforming.  What about it do you think might lead to a
Dark Age?