00:17
<JonathanNeal>
Do some of you folks work on the whatwg project?
00:17
<Hixie>
this _is_ the whatwg project
00:20
<JonathanNeal>
Nice.
00:23
<JonathanNeal>
Now we're moving to the second portion of the HTML5 adoption, which is applying it to our latest out-of-the-box theme. I have an example @ http://madison.thewikies.com/html5/liferay/
00:24
<JonathanNeal>
I'm still working on getting the code for our div based columns. So I know I'm losing everything inside the tables.
00:54
<Hixie>
o_O
00:54
<Hixie>
now leif is suggesting ARIA should be used by non-ATs
00:54
<Hixie>
going directly contrary to the point of ARIA...
01:05
<Hixie>
well this is dumb
01:05
<Hixie>
the way i designed the database API, you can't know whether you need to create the tables or not
01:06
<Hixie>
<-- dumb
01:49
<jlebar>
Hixie, does [6.11.9 History Traversal] occur on all page loads, even if I type straight into the URL bar?
01:49
<Hixie>
yes
01:49
<Hixie>
it is invoked as part of the navigation algorithm
01:49
<jlebar>
So is popstate fired on all loads?
01:49
<Hixie>
no
01:49
<Hixie>
only if there is state to pop
01:50
<jlebar>
Only if the state to pop is non-null?
01:50
<jlebar>
I could set the document pending object to null as part of the algorithm.
01:53
<Hixie>
hold on
01:55
<Hixie>
jlebar: ok, here now. looking...
01:56
<Hixie>
jlebar: popstate only fires when an entry in the session history is activated
01:56
<Hixie>
jlebar: which only occurs if the specified entry is a state object or the first entry for a Document
01:57
<Hixie>
jlebar: but yes, it fires at least once per Document load
01:57
<Hixie>
s/but/so/
01:57
<Hixie>
my statement earlier was incorrect
01:58
<Hixie>
jlebar: sorry about that!
02:01
<jlebar>
Hixie, sorry, was afk for a sec. I'm not totally clear on this: If we load index.html and then go to index.html#foo, do we get one or two popstates?
02:02
<jlebar>
Once for the initial document load...
02:02
<Hixie>
one
02:02
<jlebar>
I see.
02:02
<Hixie>
you get one for the page load
02:02
<jlebar>
And if we go back to index.html, we also don't get a popstate, because index.html doesn't have an associated state object.
02:02
<Hixie>
right
02:03
<Hixie>
er
02:03
<Hixie>
well actually...
02:03
Hixie
looks closer
02:03
<jlebar>
:)
02:03
<Hixie>
actually yes, you do
02:03
<Hixie>
you always get a popstate when traversing to the first entry in the history for a Document
02:03
jlebar
is relieved.
02:03
<jlebar>
Oh...interesting.
02:03
<Hixie>
not sure that's entirely sane
02:03
<Hixie>
but it's what it says
02:03
<jlebar>
Me either.
02:04
<jlebar>
I've been dispatching popstate whenever we dispatch onload, and also when you go back/forwards to any history entry.
02:05
<jlebar>
That might be easier for developers to deal with, since it's somewhat consistent.
02:05
<Hixie>
the original intent was to only dispatch it when hitting an entry that has been pushState()d
02:05
<Hixie>
i am not sure why we dispatch for the initial page load
02:05
<jlebar>
I think because you want to know when you go back to it?
02:05
<Hixie>
i guess, yeah
02:06
<Hixie>
yeah, that makes sense
02:06
<jlebar>
So it treats the first history entry for a document as one which was pushState()ed to.
02:06
<Hixie>
right
02:06
<Hixie>
that makes sense i guess
02:06
<jlebar>
maybe? :)
02:06
<Hixie>
:-)
02:06
<Hixie>
so the only entries for which you don't fire a popstate per spec are those that were added due to the hash changing
02:06
<Hixie>
(they get a hashchange)
02:07
<jlebar>
I see. That might be a little confusing, because clearstate does clear those.
02:07
<Hixie>
yeah, that was a recent change
02:08
<Hixie>
iirc
02:08
<Hixie>
or at least, that wasn't my original intent
02:08
<Hixie>
thing is, if we do fire popstate for everything
02:08
<Hixie>
then authors are going to have to distinguish the state=null-because-first case from the state=null-because-hashchange-will-fire-next case
02:09
<Hixie>
and that's probably harder than just dealing with those two cases
02:10
<jlebar>
Hm... I'm not yet convinced that matters. Wouldn't they listen to popstate alone in that case?
02:11
<jlebar>
* rephrased: If they just ignore hashchanges, do they care about differentiating between them?
02:11
<Hixie>
if they don't distinguish them, then each time the user changes the hash manually, the page is going to reset to the initial state
02:11
<jlebar>
Hmm
02:13
jlebar
wonders how this would change if we had a setState() function. Then you *could* associate state with the initial entry.
02:14
<Hixie>
true
02:14
<Hixie>
but it's easy enough to remember the state for the initial state
02:15
<Hixie>
i mean, what they _should_ do is if state is null, or if hashchange is called, just base the state on the url
02:15
<Hixie>
i guess we could call popstate always
02:15
<Hixie>
do any other UAs implement this yet?
02:15
<jlebar>
That's how I've been imagining it working. That way you can bookmark the page and it'll do the right thing.
02:15
<jlebar>
I'm not aware of any other UAs working on this.
02:15
<Hixie>
try e-mailing the list and seeing what others think of it
02:16
<jlebar>
Will do.
02:16
<Hixie>
if it's not too late for shipped implementations, i can change it to just always fire popstate
02:16
<Hixie>
my contribution to the web: events firing all the damn time. :-/
02:16
<jlebar>
Okay. I'll try and send out an e-mail soon. I have a few other things in there, but they don't require the Author to parse the spec for me. :)
02:16
<Hixie>
i guess nothing i've done with events is as bad as mutation events, still
02:16
<Hixie>
jlebar: k :-)
02:17
<jlebar>
Thanks, Hixie.
02:17
<Hixie>
np
02:17
Hixie
wonders why adding features to the web platform feels like a continuous exercise in damage mitigation
02:17
<Hixie>
it's like trying to build a sandcastle under a waterfall
02:18
<jlebar>
lol. And you have four large, powerful groups of people all telling you that they don't like the shape of your crenelations.
02:18
<Hixie>
fan more than four
02:19
<Hixie>
far, even
02:19
<Hixie>
there's the IETF people, the accessibility people, microsoft, google, apple, mozilla, opera, the TAG, the RDFa people...
02:19
<jlebar>
Yeah. I'm only just beginning to get into this, but...I don't envy that.
02:20
<Hixie>
eh, the sad thing is i enjoy it
02:20
<Hixie>
i'm a sucker i guess :-P
02:20
<jlebar>
:) I need to go catch my train now. Thanks again!
02:20
<Hixie>
later!
05:02
<Hixie>
shepazu: xhr was part of html5
05:02
<Hixie>
it was the first spec to be spun off
06:13
<JonathanNeal>
Hello all.
06:18
<Hixie>
anyone know how focus and blur events work on Window objects?
06:23
<othermaciej>
Hixie: I think they fire based on the window gaining or losing focus
06:23
<othermaciej>
more detail than that, I don't know
06:24
<Hixie>
i guess i'll look into it on wednesday
06:24
<Hixie>
bed now
06:24
<Hixie>
nn
06:37
<JonathanNeal>
Night, Hixie.
06:39
<JonathanNeal>
focus is when a particular element is active, like when a link is selected and has that little dotted outline in firefox, or when your mouse is pressed inside an input element and the text cursor is flashing, when you "enter" that element, that is focus.
06:39
<JonathanNeal>
Blur is the opposite and occurs only when you leave a particular element.
06:40
<JonathanNeal>
oh geez, you asked about window objects, not elements.
06:40
<JonathanNeal>
If you're talking about the window that shows you the body of the page, then it works the same as the elements.
06:41
<JonathanNeal>
when the window is active and when you're on the tab of that particular page, it's active.
06:42
<JonathanNeal>
you have to actively leave the tab by switching to another one or minimizing the window to trigger blur.
07:39
<shepazu>
Hixie: XHR came to W3C long before HTML5 did... with regards to the Rec track, it's not a "spin-off" in that sense
07:41
<annevk2>
not that long
07:41
<annevk2>
less than a year
07:42
<annevk2>
and it did came out of what became known as HTML5 later
07:49
<annevk2>
it's even part of the acknowledgments
07:51
hsivonen
thinks the maturity level req for normative references is entirely unhelpful
07:53
<annevk2>
XMLHttpRequest is not moving that fast at the moment
07:53
<annevk2>
mostly awaiting more impl feedback
07:57
<othermaciej>
shepazu: it was spun off before WHATWG Web Apps 1.0 was officially adopted as W3C HTML5
07:58
<othermaciej>
hsivonen: it somewhat makes sense to me that, if you normatively depend on an unstable spec, it's dubious to declare yourself frozen
07:58
<othermaciej>
hsivonen: but since Web technology has turned out to be a web of specs and not a hierarchy, it does seem like it could impede progress
08:01
<hsivonen>
othermaciej: implementations aren't going to stabilize at a point where they are only implementing RECs
08:02
<othermaciej>
hsivonen: that tends to undermine the idea of RECs at all
08:02
<hsivonen>
othermaciej: instead, any particular piece of the platform may advance ahead of components that rely on the service the piece provides
08:02
<othermaciej>
I am not sure what the right process is for Web standards
08:02
<hsivonen>
othermaciej: I'm not a believer in RECs, either.
08:02
<othermaciej>
implementations move too fast
08:03
<othermaciej>
for the traditional standards process
08:03
<othermaciej>
but never stabilizing anything also seems like a bad idea
08:04
<hsivonen>
it makes sense to stabilize stuff, but making spec division as granular as actual stability would be prohibitively bureaucratic
08:04
<hsivonen>
we'd have on W3C doc per DOM method
08:04
<hsivonen>
s/on/one/
08:05
<hsivonen>
even the WHATWG section-level stability isn't quite granular enough to annotate real implementation stability
08:06
<hsivonen>
I rather like the analogy Hixie made with U.S. Code
08:06
<othermaciej>
probably still more practical than doing it per-method
08:08
<annevk2>
shepazu, please...
08:09
<annevk2>
(re latest email)
08:16
<shepazu>
annevk2: what? you raised a totally pedantic point that wasn't even correct, nor was it pertinent to the subject at hand... you expected a serious reply?
08:18
<annevk2>
dream on
08:18
<hsivonen>
shepazu: what was annevk2's pedantic point?
08:20
<shepazu>
the same as hixie's, naturally... to which I'd already replied in this channel
08:27
<annevk2>
hmm
08:27
<annevk2>
the WHATWG twitter account is no longer updated?
08:28
<annevk2>
probably a problem with twitter I suppose :/
08:43
<othermaciej>
shepazu: the original contents of the W3C XHR spec were once in "Web Apps 1.0", which at the time was known unofficially as HTML5 and now is known officially as HTML5
08:44
<othermaciej>
shepazu: I'm not sure what your goal is in casting doubt on that record
09:16
<shepazu>
othermaciej: "casting doubt on that record"? I laid it out explicitly... you might as well ask what anne's point was in claiming XHR originated with the HTML5 spec... it's totally irrelevant to the explanation I was making for Shelley about what the relationship of specs on the Rec-track are
09:17
<annevk5>
I was just making a small nit about your statement being the editor of XMLHttpRequest and all
09:17
<shepazu>
why is it always a political debate when someone gets off the party line?
09:17
<othermaciej>
shepazu: the part of your email he quoted said "XmlHttpRequest (XHR) spec, which was never part of HTML5"
09:17
<othermaciej>
I think he just gave a simple factual correction, from his perspective
09:17
<othermaciej>
I don't believe anyone but you is making it political
09:17
<shepazu>
pft.
09:18
<othermaciej>
it was a side point which I don't think invalidates the rest of your comments
09:19
<aho>
void save(); // push state on state stack
09:19
<aho>
void restore(); // pop state stack and restore state
09:19
<aho>
why aren't these called push and pop?
09:19
<aho>
save and restore makes it sound like there can be only one state
09:19
<shepazu>
wtf??? http://www.fujinonbinos.com/
09:20
<shepazu>
(view source)
09:21
<hsivonen>
shepazu: awesomeness
09:21
<shepazu>
code as art?
09:22
<othermaciej>
my eyes hurt
09:22
<shepazu>
the Poseidon Series, appropriately, goes very deep
09:23
<foolip>
jgraham, Philip`: did either of you happen to implement the DOM APIs for microdata for your demos? I'm wanting to do a microdata experiment but am too lazy to implement the API myself
09:23
<annevk5>
sicking, I think another important reason for alt= has been search engines
09:24
<shepazu>
<META NAME="generator" CONTENT="Web Studio, Version 4.0 for Windows">
09:37
<jgraham>
foolip: No, mine is python based
09:38
<jgraham>
so it doesn't really have a concept of DOM
09:38
<foolip>
too bad
09:38
<jgraham>
Philip`'s was js based though. I don't know if he tried implementing the DOM APIs
09:39
<foolip>
I'll have a look then
09:40
<othermaciej>
annevk2: do search engines read alt?
09:40
<othermaciej>
annevk2: I wonder why it has not become a search engine spam vector (or maybe it has)
09:46
<annevk5>
othermaciej, I think it's in every SEO text book (not that I ever read one)
09:47
<hsivonen>
on the codec topic: did VLC regress or remove Dirac support?
10:41
<annevk5>
so Firefox does support ISO88599 unlike IE but like Chromium and Opera
10:41
<annevk5>
but it does not support ISO-8859_9 unlike IE, Chromium and Opera
10:50
<aho>
how about ISO-8859-9? :>
10:51
<annevk5>
including the question mark? </smart-ass>
10:51
<aho>
wouldnt be a question then, would it? :P
10:52
<annevk5>
you win :)
12:50
<Dashiva>
What are the non-aria uses of role now that xhtml2 is going away?
12:50
<hsivonen>
Dashiva: none
12:52
gsnedders
guesses Hixie is sleeping
12:52
<Dashiva>
Then what's up with the role vs aria-role thread?
12:54
<annevk5>
I don't think everyone is in agreement that role="" is just for WAI-ARIA
12:54
<hsivonen>
Dashiva: role is an inconsistent name when everything else in ARIA has the aria-* prefix
12:54
<hsivonen>
Dashiva: it sucks, but it's too late. not too serious to worry about, IMO.
12:55
<hsivonen>
annevk5: who disagrees except the XHTML2 WG?
12:55
<annevk5>
The people who brought it up I assume
12:56
<Dashiva>
But nobody knows here what other uses are being considered?
12:59
<annevk5>
I don't fully understand what they want
13:06
<Lachy>
I don't understand that thread at all. Jim stated in his e-mail about it that "ARIA roles are not the *only* use of the role attribute", and regardless of whether that's true or not, it refutes his own argument for making it aria specific
13:07
<hsivonen>
non-ARIA uses of role are handled by class :-)
13:08
<Dashiva>
<element element-role="div">
13:10
<jgraham>
I think he wants aria-role and role so that you can have <input type=checkbox aria-role=radio role=url>
16:13
<zcorpan>
http://www.w3.org/mid/op.uyupstgqidj3kv⊙sl
16:14
<annevk2>
back to ABNF?
16:15
<annevk2>
or XML BNF or whatever it is
16:15
<zcorpan>
it's XML 1.0 EBNF
16:16
<annevk2>
so the new rule is not matching the EBNF is ignoring the entire PI?
16:16
<zcorpan>
yeah
16:17
<zcorpan>
plus duplicate pseudo-attribute and &#x0;
16:19
<annevk2>
XML infoset? :/
16:20
<zcorpan>
the wg doesn't like having it defined in terms of the dom, because not all impl use a dom...
16:20
<zcorpan>
but they were fine with infoset
16:21
<annevk2>
hmm, it also doesn't define impl requirements anymore it seems
16:21
<annevk2>
e.g. section 5 seems more like a description for authors than implementors
16:23
<annevk2>
http://simon.html5.org/specs/xml-stylesheet5 is nicer :)
16:24
annevk2
is somewhat positively surprised with http://www.ietf.org/mail-archive/web/http-state/current/msg00191.html
16:38
<jgraham>
http://news.bbc.co.uk/2/hi/science/nature/8206280.stm
16:38
<jgraham>
MikeSmith in particular should take note of the guy's name
16:44
<Lachy>
"Professor Robert Smith? (the question mark is part of his surname and not a typographical mistake) ..." - That guy must have had some weird parents to have given him a question mark in his name
16:45
<jgraham>
Lachy: Given that it is to distinguish him from the more famous Robert Smith I guess he got his name changed
16:45
<jgraham>
Kinda like Mike(TM)Smith but moreso
16:45
<Lachy>
I hadn't read that far through the article to see that yet
16:46
<Lachy>
So I guess he added it himself
16:46
<gsnedders>
n00b
16:50
<Dashiva>
jgraham: It's a shame the study uses such a flawed model
16:51
<Dashiva>
Pretty basic mistake, too, I wonder how they missed it
16:52
<zcorpan>
gsnedders: fixed the charref issue
17:03
<jgraham>
Dashiva: I assume the study was designed to give the publicity-happy result
17:04
<jgraham>
Although that may be unduly cynical, it is worth noting that it is by someone who thought that doing a study on disease propogation using zombies was a good idea and who appended a question mark to their name
17:06
<Dashiva>
So you believe that in a model where zombies can't re-reanimate, humans and zombies could coexist peacefully?
17:08
<annevk5>
jgraham, good enough to make BBC News
17:08
<annevk5>
apparently :)
17:08
<Dashiva>
Props to BBC for finding a guy smart enough to point out the obvious flaw in the study :)
17:11
<jgraham>
annevk5: "silly season"
17:11
<annevk5>
which reminds me, where is RB?
17:13
<zcorpan>
annevk2: don't summon him please
17:15
<jgraham>
I have to say that every time Dmitry Titov posts I have a momentary panic that it is Dmitry Turin
17:18
<annevk5>
foolip, gonna blog about http://twitter.com/foolip/status/3384463111 ?
17:19
<foolip>
annevk5: maybe after a finish an experiment with the Microdata DOM API (which I have to implement first)
17:19
<foolip>
s/a/I/
17:20
<foolip>
(in JS that is, not in browser)
17:22
<gsnedders>
"If current node is a pre, textarea, or listing element, append a U+000A LINE FEED (LF) character." — does anything do that?
17:22
<gsnedders>
As far as I can tell nothing does that
17:25
<jgraham>
Doesn't seem worth worrying about unless it causes compat issues
17:25
<annevk5>
but things probably should because otherwise we lose a \n each time it's parsed and then serialized
17:26
<annevk5>
and then parsed again, doh
17:26
<annevk5>
foolip, k
17:41
<JonathanNeal>
Goodmorning!
17:41
<zcorpan>
hsivonen: a difference with parsing into the element itself when setting innerHTML (instead of doing what the spec says): <pre>.innerHTML = '\nx'
17:42
<Lachy>
Good evening JonathanNeal
17:52
<JonathanNeal>
Hello all.
17:52
<JonathanNeal>
I mean, goodevening Lachy.
17:56
<JonathanNeal>
Where are you based?
18:01
<miketaylr>
it's lunch time over here, so good afternoon JonathanNeal
18:10
<JonathanNeal>
Goodevening Lachy and goodafternoon miketaylr.
18:12
<Dashiva>
i18n('greeting')
18:12
<JonathanNeal>
Thanks for all of your guys help so far, I'm making good progress with http://madison.thewikies.com/html5/liferay/ and I'll be working on the core of the portal to have the default unstyled theme be an HTML5 semantic document.
18:49
<Lachy>
JonathanNeal, I'm in Norway
18:50
<JonathanNeal>
Groovy.
20:34
<othermaciej>
hi everyone
20:36
<TabAtkins>
Yo, Maciej
22:26
<JonathanNeal>
hi othermaciej, delayed.
22:48
<Philip`>
foolip: I tried starting to implement the microdata DOM stuff, before deciding it required more effort and knowledge than I wanted to bother with
23:04
Philip`
tries to work out how to change Google from its silly geolocated Spanish default into English
23:04
Philip`
fails, since he doesn't know the Spanish for "English"
23:05
<Philip`>
Aha, got it now, by going through Hebrew and Pirate
23:16
<annevk5>
Dashiva, took a few minutes to hack this together before going to bed: http://dump.testsuite.org/2009/encoding-matching/runtests.htm
23:18
annevk5
wonders if contentDocument works in IE8
23:21
<JonathanNeal>
FAIL
23:21
<JonathanNeal>
annevk5 doesn't like me.
23:21
<Dashiva>
annevk5: The iframe is so tiny!
23:22
<TabAtkins>
Philip: Hahaha. It's "ingles", by the way (there's an accent in there, but irrelevant here).
23:23
<annevk5>
Dashiva, the frame is not really relevant
23:24
jgraham
wishes you could tell what you failed/passed
23:24
<Dashiva>
annevk5: You might have to use contentWindow.document
23:24
<annevk5>
yeah, I thought of adding titles
23:25
<Philip`>
TabAtkins: Ah, right - I had no chance if it didn't begin with "E"
23:26
<annevk5>
now it reports the title and uses contentWindow
23:26
<Dashiva>
Hmm, or no
23:26
<TabAtkins>
Eh, as long as you can read the alphabet, most languages translate English relatively understandably. It *does* mean you may have to read through a whole list of unknown foreign words, though.
23:26
<Dashiva>
That wasn't the problem, it's the onload that doesn't fire
23:27
<annevk5>
ok, I'll change some things
23:27
<Dashiva>
Apparently it doesn't support onload, only attachEvent with load?
23:30
<Dashiva>
Or you can use the onload attribute
23:31
<annevk5>
try it now?
23:33
<annevk5>
can't believe this is taking more than 30 minutes
23:34
<Dashiva>
Yeah, it's running now
23:35
<annevk5>
cool
23:35
<Dashiva>
Could you include a \n in the output? :)
23:35
<annevk5>
you mean \r then?
23:35
<Dashiva>
Oh, right. Yeah.
23:36
<TabAtkins>
God, the /r/n debacle is one of the most nonsensical interop failures ever.
23:36
<annevk5>
done
23:36
<annevk5>
can't believe IE still hasn't fixed that
23:36
<annevk5>
well I can
23:36
<annevk5>
it just annoys me
23:37
<Dashiva>
WTf... it's still all on one line
23:38
<annevk5>
ok, i'll fix that too before I go to sleep
23:38
<annevk5>
hold on
23:38
<Dashiva>
They don't appear if I alert document.body.innerHTML either
23:40
<annevk5>
reload
23:41
<Dashiva>
works
23:41
<Dashiva>
Now it just needs color coding in the list and you're done ;P
23:41
<annevk5>
mwaha, guess what finger I'm holding up :p
23:41
<annevk5>
anyway, thanks for the help, much appreciated :)
23:42
<Dashiva>
25-27 all fail, by the way
23:42
<annevk5>
ah interesting
23:42
<annevk5>
I wonder why the other browsers recognize those aliases
23:42
<Dashiva>
Oh, there's a bug
23:42
<Dashiva>
...
23:42
<annevk5>
...
23:42
<Dashiva>
No, that's just me needing sleep
23:42
<annevk5>
profit
23:43
<TabAtkins>
I get a fail on 25 and 27 on FF 3.5.2
23:43
<annevk5>
me too
23:43
<Dashiva>
Opera fails 001... wtf
23:43
<annevk5>
TabAtkins, yeah me too, not entirely unexpected, with the last few I was just playing around
23:43
<TabAtkins>
Plenty of others, too. But just saying that "the other browsers" don't recognize those.
23:43
<TabAtkins>
Ah, kk.
23:43
<annevk5>
Dashiva, we don't alias ISO-8859-9 to Windows-1254 yet
23:43
<annevk5>
filed a bug on it today
23:43
<Dashiva>
Opera passes 25-27, though. Heh.
23:44
<annevk5>
Firefox doesn't do that either, hence it failing tests
23:44
<annevk5>
if it did that it would pass mosts
23:45
<annevk5>
anyway, nn
23:45
<Dashiva>
nn
23:46
<TabAtkins>
nn = nighty night, right?
23:47
<Dashiva>
never normalize
23:47
<TabAtkins>
Ah, makes sense.
23:48
<TabAtkins>
wtf, php. It's telling me that it's max int is 2^31, as expected, and that is uses 4-bit ints. But it's also storing and showing me the number 11billion, which should require 34 bits to store.
23:49
<mpilgrim>
annevk5: are you looking for results of those encoding tests?
23:49
<mpilgrim>
if so, which browsers?
23:50
<Dashiva>
I think he has all he needs for now
23:50
<mpilgrim>
ok
23:50
<mpilgrim>
what do they test?
23:50
<Dashiva>
charset parsing and aliasing, as far as I know
23:51
<TabAtkins>
That's what it looks like, yah.
23:51
<Dashiva>
mpilgrim: When will the first chapter of your new book be ready? :)
23:51
<gsnedders_>
TabAtkins: It implicitly converts from ints to floats as needed
23:51
<mpilgrim>
as soon as i finish the last book
23:52
Dashiva
chuckles at 4-bit ints
23:52
gsnedders_
noticed that too :)
23:54
<TabAtkins>
Hrm. Then how is it keeping accuracy? Just checked against my lisp implementation the value of a 7-digit number in base 32 (relevant to what I'm doing), and PHP and Lisp agree.
23:54
<TabAtkins>
And I know Lisp uses bigints.
23:54
<TabAtkins>
(the max 7-digit value, that is)
23:55
<TabAtkins>
I think my PHP is secretly using 8B ints and lying to me.
23:55
TabAtkins
just got what Dashiva and gsnedders were chuckling at.
23:55
<Dashiva>
mpilgrim: You note that Opera doesn't support text in ch4, but one of the examples earlier in the chapter uses text (and thus doesn't match the text in Opera)
23:55
<mpilgrim>
yes
23:55
<mpilgrim>
the print edition will obviously use static images
23:56
<mpilgrim>
i tried to be up-front about it in each section
23:56
<mpilgrim>
if your browser has an "x", the section examples may not make sense
23:56
<mpilgrim>
might end up using static images for the online edition too, and having separate example pages
23:56
<Dashiva>
*nod*
23:57
<Dashiva>
Just thought this might be confusing since the x for text won't appear for another few scrolls