03:35
<Von_Davidicus>
So what, precisely, are the advantages of upgrading a page from HTML 4.01 to HTML 5?
04:14
<Hixie>
Von_Davidicus: what do you mean by "upgrading a page from HTML 4.01 to HTML 5"?
04:15
<Von_Davidicus>
As in changing the Document Type Declaration to <!DOCTYPE html> and implementing the needed changes to make it all valid and--if need be--fixing the JavaScript so that it works with the resultant code.
04:16
<Hixie>
there's no advantage to changing the doctype line unless you are changing it from something that used to trigger quirks mode
04:16
<Hixie>
if that is the case, the advantage is slightly saner browser behaviour in weird cases
04:17
<Von_Davidicus>
I don't think it triggered quirks mode.
04:17
<Hixie>
if it's not valid contemporary HTML then the advantage to making it so is you avoid whatever problems are the reason for us to make it invalid
04:17
<Hixie>
but if the page is working fine, there's no reason to do make work
04:18
<Von_Davidicus>
Well, according to the HTML5 validator, I was abusing the tar out of the headers element.
04:18
<Hixie>
(most HTML4 docs should still be valid today if they were strict mode docs and didn't do anything problematic)
04:18
<Hixie>
<headers>?
04:18
<Von_Davidicus>
er...
04:18
<Von_Davidicus>
headers attribute.
04:18
<Hixie>
aah
04:18
<Hixie>
yeah, the validator is much better now than validators in the html4 days were
04:18
<Von_Davidicus>
They weren't referring to an ID set in the table, but to IDs of input elements.
04:18
<Hixie>
hah
04:19
<Hixie>
that is indeed the kind of thing the modern validators will catch if they're following the spec :-)
04:20
<Von_Davidicus>
I kinda like the [data-*] attribute--that's what I used to replace what I was using the headers attributes for (I came up with [data-keywords])
04:22
<Von_Davidicus>
Is there a way to say "this data-* attribute should work like an IDREFS attribute"?
04:26
<Von_Davidicus>
One thing I liked about using the headers attribute was if I had a keyword that was misspelled or something (which meant the script wouldn't work), it showed up as a warning. The relevant page is http://www.mrinitialman.com/Furcadia/DragonSpeakLines.shtml
04:36
<Hixie>
Von_Davidicus: data-* is just useful for scripts, so the way you make it work like an idrefs attribute is to make the script treat it like an idref attribute
04:37
<Hixie>
Von_Davidicus: if you want warnings, just have your script check that the values are right and call console.warning() if they're not
04:53
<Von_Davidicus>
Okay.
04:57
<Von_Davidicus>
By the way, with the way you phrased your comment, you made it sound like data-* was created mainly for scripts. Is that true? :)
05:18
<Hixie>
only for scripts, pretty much
05:18
<Von_Davidicus>
*Nod.*
05:36
<Von_Davidicus>
I'll admit--I've done some pages where an upgrade from HTML 2.0 wouldn't change much.
05:37
<Hixie>
i know the feeling :-)
05:38
<Von_Davidicus>
Oh, and I've figured out something: I've finally discovered that XSLT is great when it's used as a client-side processing language (such as referencing multiple XML files or creating an SVG diagram out of an XHTML document). How does it hold up against, say, a *server*-side processing language? Aaaaaaahm....
05:38
<Von_Davidicus>
So, it's got advantages over XHTML.
05:39
Hixie
is skeptical that XSLT is great anywhere, but ok
05:39
<Von_Davidicus>
The only Server-Side language XSLT can outperform? Server-side includes. And I'm not certain about that, either.
05:41
Von_Davidicus
dusts off SSI from time to time--such as when a page is SO big that it's a pain to update.
07:20
<Von_Davidicus>
Thanks, Hixie. That script-based solution works well.
07:20
<Hixie>
np
07:21
<Von_Davidicus>
I think I've realized why I use XSLT.
07:21
<Von_Davidicus>
I think I do it to show that I *can* pull it off.
07:22
<Hixie>
that's a fine reason to solve a rubix cube
07:23
<Hixie>
not necessarily a good reason to use a programming language :-P
07:24
<Von_Davidicus>
Maybe not. But I'd hate to let something I learned go comepletely to waste. And I *have* used webhosts before that didn't have server-side capabilities.
07:25
<Hixie>
heh
07:25
<Hixie>
i recommend reading up on something called the "sunk cost fallacy"
07:25
<Von_Davidicus>
*glares at Furnation*
07:28
<Von_Davidicus>
Furnation is/was (these days, mostly 'was'), a *very* well-known web host within a certain fandom, and it was one of those hosts that ran on donations as well as their own publishing company. The latter was a good idea and might have carried the host better if they hadn't dropped the ball so often they left a permanent dent on the floor.
07:29
<Von_Davidicus>
But it had no server-side capabilities. Not even includes.
07:30
<Von_Davidicus>
So I suppose in *that* situation, it would be either hardcode everything, use XSLT, or use frames.
07:47
<Von_Davidicus>
Not sure what would be considered the bigger coding sin these days: XSLT or frames.
07:56
<Ms2ger>
XSLT
08:09
<AryehGregor>
Von_Davidicus, you could build it all using AJAX!
08:09
<AryehGregor>
Then at least if someone else needed to take over maintaining it, maybe they would understand your code . . .
08:11
<Von_Davidicus>
AJAX is based on JavaScript. I'm not sure which would be worse, trying to understand someone else's XSLT or trying to understand someone else's JavaScript.
08:11
<AryehGregor>
A lot of people actually know JavaScript.
08:11
<AryehGregor>
Nobody except you and four other people knows XSLT, or indeed has ever heard of it.
08:12
<AryehGregor>
The other four people are members of the XSLT WG.
08:13
<AryehGregor>
(I'm being mean -- I'm sure it's better-known than lots of other niche technologies. In #whatwg we're just bitter about it because it's one of the technologies *we* have to deal with and it just makes our lives more difficult.)
08:13
<AryehGregor>
(Slightly.)
08:14
<Von_Davidicus>
why does it make your lives more difficult?
08:15
<AryehGregor>
Because browsers still support it.
08:15
<AryehGregor>
And it's fairly large and complicated.
08:15
<AryehGregor>
The web is complicated enough.
08:15
<AryehGregor>
We don't need more things that are complicated but that almost no one uses.
08:18
<Von_Davidicus>
How does that get in the way of HTML5 development?
08:18
<AryehGregor>
It doesn't get in the way of it, much. But we're trying to define the web platform as a whole, so things that aren't worth the effort to define and implement properly are things we'd prefer just go away.
08:19
<Von_Davidicus>
Oh, okay.
08:19
<AryehGregor>
XSLT is marginal and self-contained enough that I don't think we hate it too much, though.
08:19
<AryehGregor>
Unlike, say, quirks mode.
08:20
<Von_Davidicus>
I just remembered something: Doesn't AJAX require the *server* to support it? Because I think I tried AJAX on Furnation and it didn't work.
08:20
<AryehGregor>
No, I'm pretty sure not.
08:20
<AryehGregor>
It's just an async HTTP request. (Or sync, as the case may be.)
08:21
Von_Davidicus
looks at his Coach Random Website, and wonders if it's worth it to switch it over to HTML5.
08:21
<Ms2ger>
Don't speak of sync xhr ;)
08:23
<Von_Davidicus>
Why not?
08:33
<Von_Davidicus>
What would you guys say?
08:34
<Von_Davidicus>
Currently, it's an XML+XSLT+XSD website.
08:39
<AryehGregor>
Sync XHR murders performance. It causes the page to freeze until the request completes.
08:39
<AryehGregor>
These days we never ever expose sync APIs to regular pages unless they're guaranteed to return very quickly.
08:39
<AryehGregor>
Anything else is async, like IDB.
08:39
<AryehGregor>
localStorage is similar.
08:39
<AryehGregor>
We expose sync APIs to workers, though.
08:48
<Von_Davidicus>
Hello, annevk.
08:49
<annevk>
hey
08:50
<Von_Davidicus>
Question: should I take an old site of mine that's written in XML+XSLT+XSD and redo it as HTML5?
08:54
<annevk>
heh, that's kind of hard for me to answer :)
08:54
<Philip`>
Von_Davidicus: Yes - it might not provide any benefit to the site itself if it's already working fine, but it might cure you of your constant desire to use XSLT
08:54
<Von_Davidicus>
It *is* online.
08:57
AryehGregor
concurs with Philip`
09:13
<Von_Davidicus>
So, Philip? Instead of using XSLT as a client-side language, I should be using COBOL for server-side?
09:24
<Von_Davidicus>
Question about HTML5. Is it possible to have one list reference another as being a continuation of that list?
09:25
<Ms2ger>
No
09:25
<Ms2ger>
You can use the start attribute to fix up the numbering
09:25
<Von_Davidicus>
Oh, the start attribute is back?
09:26
<Von_Davidicus>
I know HTML 4.01 Strict didn't allow it.
09:26
<Ms2ger>
It is
09:29
<Von_Davidicus>
Only time I'd ever used it was to see what Lynx would do with an <ol> that went over the numbering limit.
09:31
<Ms2ger>
What did it do?
09:33
<Von_Davidicus>
It went from 2,147,483,647 to -2,147,483,648
09:34
<Von_Davidicus>
That being said, it's hard to imagine an ordered list so long that it would surpass the limit for the roman numeral styles.
09:38
<Von_Davidicus>
The highest they can show is MMMCMXCIX or mmmcmxcix (depending on whether you're using upper or lower case). That's 3,999 <li> elements (unless you're playing with the count), which gets into the realm of "can the average computer handle this webpage". :)
15:01
<Ms2ger>
"The jQuery forum appears to be completely dysfunctional without JavaScript enabled, which is sad"
16:27
gsnedders
wonders about optional static typing for ES again, and decides it's probably better not to think about
16:28
<Ms2ger>
Correct
16:28
<gsnedders>
So, SpiderMonkey is impl ES6 by default, V8 is impl ES6 behind an experimental pref…
16:31
gsnedders
wonders
16:33
<Ms2ger>
We really should have implemented Number.mozIsNaN...
16:36
<gsnedders>
Ms2ger: Yes, prefixes are definitely better!
16:41
<gsnedders>
Heh, broken 10k unread emails on www-style.
16:41
<gsnedders>
I really should just leave the WG.
16:57
<Hixie>
i wish addHitRegion() would be implemented already :-)
17:44
<Hixie>
is there a spec for MouseEvent?
17:45
<Hixie>
(for UAs firing events using it, that is, not for the interface)
17:46
<Hixie>
in particular i'm interested in what said spec says to set offsetX to
17:46
<Hixie>
hm, looks like http://dev.w3.org/csswg/cssom-view/#extensions-to-the-mouseevent-interface does it
17:46
<Hixie>
wonder how accurate that is
17:49
<Ms2ger>
Hixie, only one way to find out... Write tests ;)
17:50
<Hixie>
kinda am
17:50
<Hixie>
(i'm writing code that uses it, we'll see if it works!)
18:02
<Hixie>
aw man
18:02
<Hixie>
want. addHitRegion().
18:44
<jgraham>
Hixie: If writing code that used something counted as writing tests, the web would be well tested and have great interoperability
18:55
<asmodai>
I guess most of you saw http://vimeo.com/43380467 ?
18:56
<Ms2ger>
I've heard of it
18:57
<Ms2ger>
hsivonen debunked some points at http://wiki.whatwg.org/wiki/Bad_Ideas
19:00
<Hixie>
people fall into three camps as far as i can see:
19:00
<Hixie>
1. people who don't realise the web sucks
19:01
<Hixie>
2. people who realise the web sucks but don't understand why it is so wildly successful and so make silly suggestions for improving it that themselves fail
19:01
<Hixie>
3. people who write web standards and go home and cry every night
19:02
<Hixie>
the guy in that video falls very clearly into #2: he doesn't understand what he's criticising
19:03
<Hixie>
(e.g. he thinks the w3c had anything to do with canvas having .arc(), he doesn't understand html's version history, he doesn't understand how paragraphs are specced, etc)
19:03
asmodai
pats Hixie
19:03
<asmodai>
Hixie: Hope you don't cry too much every night
19:08
<asmodai>
I personally think he does have some valid points on weird behaviour and some things not being straightforward.
19:08
<Hixie>
yes
19:08
<Hixie>
the web sucks
19:09
<asmodai>
And in some other areas he really generalises.
19:09
<Hixie>
nobody argues that
19:09
<asmodai>
Heh
19:14
<jarek>
why there are no b-splines in SVG?
19:15
<jarek>
were they considered not useful enough? Is there an easy way to emulate them with beziers?
19:15
<Ms2ger>
Because W3C
19:15
Ms2ger
runs
19:15
<jarek>
Silverlight has b-splines...
19:16
<Hixie>
you may have better luck getting a real answer from #svg on w3c's irc network
19:17
<Hixie>
since it looks like none of us know :-)
19:17
<Hixie>
canvas doesn't have b-splines because apple didn't add them when they invented canvas and nobody has made a convincing argument for them since
19:17
<Hixie>
but i dunno about svg
19:18
<asmodai>
Ms2ger: lol
19:20
<spobat>
:)
19:20
<Ms2ger>
Good morning
19:20
<Philip`>
Canvas should support NURBS, because NURBS is a great acronym
19:21
<Ms2ger>
Philip`, only after you fix all the canvas test bugs :)
19:21
<spobat>
non uniform rational b splines?
19:21
<Ms2ger>
Apparently so
20:46
<gsnedders>
Hixie: What's the reason not to have a circle API in canvas, BTW? "API usability" not a good enough reason? :)
20:48
<Hixie>
what would it do?
20:48
<Hixie>
add a circle to the path? fill a circle?
20:50
<gsnedders>
Hixie: The same as drawArc?
20:51
<Hixie>
assuming you mean the same as "arc()" (there's no drawArc), then what's the point of having two functions that do the same thing
20:56
<gsnedders>
Hixie: API usability.
20:56
<gsnedders>
arc() isn't obviously what you want for a circle.
20:56
<Hixie>
API usability is as much helped by adding convenience functions as it is hurt by having a large surface area
20:56
<gsnedders>
Sure, there's no *technical* advantage.
20:57
<Hixie>
is ellipse() obviously what you want for a circle?
23:12
<gsnedders>
Heh, so easy to make Erik Arvidsson's email mean Google's webapps aren't on the "open web".
23:12
<gsnedders>
(Email to es-discuss, this is)
23:12
<gsnedders>
There again, given they have UA sniffing to start with on many things, that's rather self-evident.