ACM Queue has an interview with one of the people behind XML:
http://www.acmqueue.org/modules.php?name=Content&pa=showpage&pid=282
Tim Bray says:
"Just like anybody who has ever used Lisp who then learned markup,
when I first learned SGML I was mostly against attributes. I just
couldn�t see why you needed two different ways to mark up data
structures. Eventually, I observed that they didn�t really add
that much to the difficulty of processing. I also observed that
there�s something in the human mind that likes having these two
markup idioms to use.
I eventually decided that they were not harmful and people liked
them, so why go against them? Consider �<a href=�somewhere�>a
label</a>�. That just feels right; that just feels like a good
idiomatic succinct piece of design to me, and it uses both
elements and attributes."
Have you also noticed that "the human mind" clearly prefers
<a href="target">some text</a> to
(a (href "target") "Some text") ?
The first one just feels right. That must be because it uses both
attributes and elements for no apparent reason.
Ulrich Hobelmann wrote:
> Have you also noticed that "the human mind" clearly prefers
> <a href="target">some text</a> to
> (a (href "target") "Some text") ?
>
> The first one just feels right. That must be because it uses both
> attributes and elements for no apparent reason.
Well, exactly, it's his mind... or maybe I ain't human.
But the example is flawed. It feels okay IMHO because it's actually an
example from HTML and is real markup*, as opposed to the usual
XML-as-a-tree-data-structure which is what people who know sexps
usually mean "XML is just crappy sexps".
*virtual highlighter pen over underlying data stream, saying "the
a-highlighted phrase in the stream should have the link target
associated with it".
Contrast (n.b. not supposed to be an example of a _good_ xml
structure for the application domain)
<ticket>
<price currency="euro">90</price>
<reservation>
<hotel>spendalot</hotel>
<booking type="special">134-5a</booking>
</reservation>
</ticket>
(ticket
(price (currency "euro") 90)
(reservation
(hotel "spendalot")
(booking (type "special") "135-5a")))
Real Markup vs. trees is one of the reasons I dislike XHTML: they ruined
a reasonably good people-friendly permissive best-effort-interpreted
markup language by insisting on "proper nesting" to make a
computer-friendly tree structure.
To make my woolly point clearer, consider a made-up system of a markup
language (not saying it's a good one!)
(doc
(meta (108 (link :ref "target")
(109 (bold))
(content "This [109]is [108]some[/109] text[/108]"))
That's marking up an underlying single-level content stream, not
specifying a tree data structure.
In pure-markup HTMLoid style:
This <b>is <a href="target">some</b> text</a>
Nowadays, you're supposed to write:
This <b>is </b><a href="target"><b>some</b> text</a>
Which is CRAP! (for humans manually writing html).
Most of the W3C's "innovations" seem to be in making it harder
for humans to slap together a web page in a few minutes.
David Golden <············@oceanfree.net> writes:
> Real Markup vs. trees is one of the reasons I dislike XHTML: they ruined
> a reasonably good people-friendly permissive best-effort-interpreted
> markup language by insisting on "proper nesting" to make a
> computer-friendly tree structure.
I disagree with this point. I've been using XHTML 1.0 Strict and
typing it in by hand using Emacs. It's not any more work than the old
HTML 2.0 type whatever you want and see what the browser does with
it. I like that I can type markup that passes the validator and looks
like what I want it to in my browser.
My past experience is that "best-effort-interpreted" has been a poor
effort with wildly different results in different browsers. And I
don't do complicated stuff (I'm not a web designer).
--
An ideal world is left as an excercise to the reader.
--- Paul Graham, On Lisp 8.1
David Steuber wrote:
> My past experience is that "best-effort-interpreted" has been a poor
> effort with wildly different results in different browsers. And I
> don't do complicated stuff (I'm not a web designer).
>
Well, fair enough, lots of people want picture-perfect. Heck, my
site is total form over function that wouldn't be possible
without some layout precision (as it is my concessions to IE
horribleness make it pretty kludgy and I should update it to be
more compliant and just include a "get firefox" button under a
"why does this site look crap in IE" FAQ)
But I think they (yeah, they, those people!) could have allowed precise
document layout, perhaps more easily even, without requiring
shoehorning everything into a tree. Yes, you can pretty much always
make something look like a tree (and boy, does that happen a lot in
lisp, and compsci in general), but...
The sheer _dishonesty_ of all those XML people just boggles the mind.
Are they really that stupid, or are they simply pretending to not get
it? I have to assume the latter; otherwise there's even less hope for
humanity.
For example, they say that S-expressions are bad because if you have
something like "))))))))))" at the end of one, and you try to hack the
expression in some crappy editor like Notepad, it's hard to tell
whether you have the right number of parens. But if, on a separate
occasion, someone complains about having to type all those incredibly
verbose end-tags (even when the start- and end-tags are 5 characters
apart), the reply is invariably: "Why aren't you using a _real_
editor, with tag-completion?"
Which prompts the question of why, if a Real Editor is assumed, do we
need the idiotic end-tags in the file to begin with? There's an
obvious contradiction here, and I'm not sure whether people such as
Bray and Prescod realize it (surely they're not that clueless?). At
any rate, these self-appointed big-shots are the ones getting quoted
and worshipped and re-quoted, while the obvious questions go unasked.
It's sickening.
David Mullen wrote:
> The sheer _dishonesty_ of all those XML people just boggles the mind.
> Are they really that stupid, or are they simply pretending to not get
> it? I have to assume the latter; otherwise there's even less hope for
> humanity.
Yes, you are right. OTOH I like to think they just don't know better,
so I don't have to assume they are actually evil ;)
> For example, they say that S-expressions are bad because if you have
> something like "))))))))))" at the end of one, and you try to hack the
> expression in some crappy editor like Notepad, it's hard to tell
> whether you have the right number of parens. But if, on a separate
> occasion, someone complains about having to type all those incredibly
> verbose end-tags (even when the start- and end-tags are 5 characters
> apart), the reply is invariably: "Why aren't you using a _real_
> editor, with tag-completion?"
The irony is that there are HUGE XML-editors out there, ugly programs
like XML-spy (I still haven't understood what it's good for). Obviously
a language like sexps that can have a parser in a couple of kilobytes
and a simple, lightweight editor is much less powerful than one that
requires powerful parsers that distinguish between elements and
attributes and that is the basis for a growing industry of tools like
XML-spy ;)
And )))))) is a lot better than
</foo>
</bar>
</quux>
</blarp>
</qeuelct>
</the-real-end>
at the end of a structure. But it takes less space, less memory, maybe
the file needs no zipping, because it's only almost half as big, so
people need less memory -> destroys jobs. But isn't that our job as CS
people anyway? *evil-laugh*
> Which prompts the question of why, if a Real Editor is assumed, do we
> need the idiotic end-tags in the file to begin with? There's an
> obvious contradiction here, and I'm not sure whether people such as
> Bray and Prescod realize it (surely they're not that clueless?). At
> any rate, these self-appointed big-shots are the ones getting quoted
> and worshipped and re-quoted, while the obvious questions go unasked.
> It's sickening.
They aren't self-appointed I think. At least Mr Bray seems quite
modest. I think people are just too stupid. They need (or want)
heroes, people that create "state-of-the-art technology", that are on
the cutting edge of things, be it XML, J2EE, PHP, or what else is en
vogue right now. Celebrities. It's like Britney Spears or Brad Pitt
for CS people :D
> They aren't self-appointed I think. At least Mr Bray seems quite
> modest. I think people are just too stupid. They need (or want)
> heroes, people that create "state-of-the-art technology", that are on
> the cutting edge of things, be it XML, J2EE, PHP, or what else is en
> vogue right now. Celebrities. It's like Britney Spears or Brad Pitt
> for CS people :D
Heh. Of course, I don't mean to imply that Tim Bray's only goal is
his own popularity; if that were true he'd have no time to program at
all. But he clearly has time to create things, albeit
less-than-tasteful things. My point was that all of these appeals to
"The People" sound an awful lot like hero-posturing to me. And I'm
suspicious of those who seem a bit too concerned with how The People
feel about parentheses, etc., and presume to speak for The People; I
have to question the purity of their motives.
David Mullen wrote:
>>They aren't self-appointed I think. At least Mr Bray seems quite
>>modest. I think people are just too stupid. They need (or want)
>>heroes, people that create "state-of-the-art technology", that are on
>>the cutting edge of things, be it XML, J2EE, PHP, or what else is en
>>vogue right now. Celebrities. It's like Britney Spears or Brad Pitt
>>for CS people :D
>
> Heh. Of course, I don't mean to imply that Tim Bray's only goal is
> his own popularity; if that were true he'd have no time to program at
> all. But he clearly has time to create things, albeit
> less-than-tasteful things. My point was that all of these appeals to
> "The People" sound an awful lot like hero-posturing to me. And I'm
> suspicious of those who seem a bit too concerned with how The People
> feel about parentheses, etc., and presume to speak for The People; I
> have to question the purity of their motives.
Let's be fair. In the quoted statement, Tim Bray doesn't appeal to "the
people" like you suggest.
Pascal
Ulrich Hobelmann wrote:
> And )))))) is a lot better than
> </foo>
> </bar>
> </quux>
> </blarp>
> </qeuelct>
> </the-real-end>
> at the end of a structure.
You have to take the background of the target audience into account.
They are used to this:
}
}
}
}
}
}
;)
Pascal
Pascal Costanza <··@p-cos.net> writes:
> You have to take the background of the target audience into account.
> They are used to this:
>
>
> }
> }
> }
> }
> }
> }
>
>
> ;)
^^
So is this last thing some sort of bi-graph representation of the close
curly bracket "}" ?
--
Thomas A. Russ, USC/Information Sciences Institute
Thomas A. Russ wrote:
> Pascal Costanza <··@p-cos.net> writes:
>
>>You have to take the background of the target audience into account.
>>They are used to this:
>>
>>
>> }
>> }
>> }
>> }
>> }
>>}
>>
>>
>>;)
>
> ^^
>
> So is this last thing some sort of bi-graph representation of the close
> curly bracket "}" ?
Probably, but this is only a conjecture that they couldn't prove yet.
However, there will be a paper in a conference soon that illustrates a
static type system for a limited subset.
Pascal
Pascal Costanza <··@p-cos.net> wrote:
+---------------
| Thomas A. Russ wrote:
| > Pascal Costanza <··@p-cos.net> writes:
| >>;)
| >
| > ^^
| > So is this last thing some sort of bi-graph representation of the close
| > curly bracket "}" ?
|
| Probably, but this is only a conjecture that they couldn't prove yet.
| However, there will be a paper in a conference soon that illustrates a
| static type system for a limited subset.
+---------------
If it hasn't already been done. See:
<http://www.vault.com/surveys/email_behavior/email_behavior.jsp?results=19>
[part of <http://www.vault.com/surveys/email_behavior/email_behavior.jsp>]
<http://www.gvu.gatech.edu/user_surveys/survey-1998-10/graphs/technology/q88.htm>
[But no mention of smiley faces!! A serious lack!]
-Rob
-----
Rob Warnock <····@rpw3.org>
627 26th Avenue <URL:http://rpw3.org/>
San Mateo, CA 94403 (650)572-2607
On 4218 September 1993, Ulrich Hobelmann wrote:
> The irony is that there are HUGE XML-editors out there, ugly programs
> like XML-spy (I still haven't understood what it's good for).
That's a simple one: there are a lot of people who believe that they
really do need *specialized* state-of-the-art tools for whatever job,
regardless of whether there are some general tools that would do the
job equally well. Hey, that's marketing: there have to be some unique
selling points. cll-people are not really their target, I would guess.
> Obviously a language like sexps that can have a parser in a couple of
> kilobytes and a simple, lightweight editor is much less powerful than
> one that requires powerful parsers that distinguish between elements
> and attributes and that is the basis for a growing industry of tools
> like XML-spy ;)
Right. You got the marketing idea.
> And )))))) is a lot better than
> </foo>
> </bar>
> </quux>
> </blarp>
> </qeuelct>
> </the-real-end>
> at the end of a structure.
Why is that? From a memory or processing point of view? That's
obviously not the point (of view) of Bray.
> They aren't self-appointed I think. At least Mr Bray seems quite
> modest. I think people are just too stupid.
I beg to differ. I think a lot of people just don't have the same
level of knowledge (especially when compared with the readers of cll).
Some people even have problems with structure when using some
specialized editor that makes the skeleton of their documents so
obvious, it's a wonder they're not hiding in some corner trembling
with horror. I think, it's all about catering to such people.
Holger
--
--- http://www.coling.uni-freiburg.de/~schauer/ ---
"It's a bondage-and-discipline issue -- this is *my* class, I made it
and *you* aren't going to corrupt its precious bodily fluids by
defining your filthy methods on it, oh no."
-- Tim Bradshaw in comp.lang.lisp
David Mullen wrote:
> The sheer _dishonesty_ of all those XML people just boggles the mind.
> Are they really that stupid, or are they simply pretending to not get
> it? I have to assume the latter; otherwise there's even less hope for
> humanity.
It's herd behavior, but I've been seeing some signs of backlash. Reality
does set in eventually.
Paul
Paul F. Dietz wrote:
> David Mullen wrote:
>
>> The sheer _dishonesty_ of all those XML people just boggles the mind.
>> Are they really that stupid, or are they simply pretending to not get
>> it? I have to assume the latter; otherwise there's even less hope for
>> humanity.
>
> It's herd behavior, but I've been seeing some signs of backlash. Reality
> does set in eventually.
This sounds interesting - could you be more specific?
Pascal
Pascal Costanza wrote:
>> It's herd behavior, but I've been seeing some signs of backlash. Reality
>> does set in eventually.
>
> This sounds interesting - could you be more specific?
Well, I'm hearing stories of people trying to use XML and suffering terrible
performance hits.
And there was a conference call the other day, where someone suggested that
some file be redone in XML, causing general laughter.
Paul
"Paul F. Dietz" <·····@dls.net> writes:
> Well, I'm hearing stories of people trying to use XML and suffering
> terrible performance hits.
To the point where there are people looking at creating a binary
version of XML. Of course I would think ASN.1 would be there for that
(or even RPC / XDR to some extent).
--
David Magda <dmagda at ee.ryerson.ca>, http://www.magda.ca/
Because the innovator has for enemies all those who have done well under
the old conditions, and lukewarm defenders in those who may do well
under the new. -- Niccolo Machiavelli, _The Prince_, Chapter VI
David Magda <··················@ee.ryerson.ca> writes:
> "Paul F. Dietz" <·····@dls.net> writes:
>
>> Well, I'm hearing stories of people trying to use XML and suffering
>> terrible performance hits.
>
> To the point where there are people looking at creating a binary
> version of XML. Of course I would think ASN.1 would be there for that
> (or even RPC / XDR to some extent).
Careful what you wish for...
<http://java.sun.com/developer/technicalArticles/WebServices/fastWS/>
As far as I understand it, X.694 maps between the standard (but widely
disliked) XSD and ASN.1, X.691 provides a way to encode and decode
ASN.1 (ASN.1 provides a way to describe data in the abstract, and
there are several standard encodings including BER (used in LDAP), its
subset DER (used in X.509 for certificates, for example), XER (not
much used, as far as I know, an XML encoding), and X.691's PER, a
compact encoding).
The idea behind fastWS is presumably obvious to anyone: represent XML
Infosets using PER, by converting your favourite XML schema to ASN.1,
and then that provides a way to encode particular XML entities.
I've no idea whether anybody actually uses this for anything.
David Mullen wrote:
> For example, they say that S-expressions are bad because if you have
> something like "))))))))))" at the end of one, and you try to hack
the
> expression in some crappy editor like Notepad, it's hard to tell
> whether you have the right number of parens. But if, on a separate
> occasion, someone complains about having to type all those incredibly
> verbose end-tags (even when the start- and end-tags are 5 characters
> apart), the reply is invariably: "Why aren't you using a _real_
> editor, with tag-completion?"
>
> Which prompts the question of why, if a Real Editor is assumed, do we
> need the idiotic end-tags in the file to begin with? There's an
> obvious contradiction here, and I'm not sure whether people such as
> Bray and Prescod realize it (surely they're not that clueless?). At
> any rate, these self-appointed big-shots are the ones getting quoted
> and worshipped and re-quoted, while the obvious questions go unasked.
> It's sickening.
So everyone else's code doesn't look like the following?!
(loop for i in '(1 2 3 4 5);</1>
;</quote>
sum (1+ i);</1+>
);</loop>
Lisp can co-opt the disadvantages of pretty much anything.
Tim Bray's assertion of "Lisp aesthetics" was odd though. Similar
arguments have been advanced against LOOP and keyword arguments. I
suspect aesthetics isn't at all the reason why people dislike
attributes in XML, even if they're nice in HTML.
Maybe he means he had a Scheme background, where aesthetics seems to be
a greater consideration?
Ulrich Hobelmann wrote:
> Have you also noticed that "the human mind" clearly prefers
> <a href="target">some text</a> to
> (a (href "target") "Some text") ?
>
> The first one just feels right. That must be because it uses both
> attributes and elements for no apparent reason.
This is a toy example and doesn't tell us a lot. There is a clear
tendency in XML to put opening tags and closing tags on single lines,
whereas with s-expressions closing parentheses are almost always lined
up together in a single line, and not even on a line of their own. The
net effect is that XML consumes a lot more vertical space than
s-expressions. This hinders understandability of complex structures: The
more you can see on a single screen, the better overview you have about
what you are reading, the more likely it is that you understand what is
going on.
Pascal