Tavern on the Meta

drachenstern

9:00 PM

I think he's talking about inlining images, or you could always use sprited images

yep, inline images

Moshe

@drachenstern I was talking aobut rchern's smiling habits and getting ppl to understand her brain

Wes

@DanGrossman content images shouldn't be specified in CSS

data urls don't work in IE6

Moshe

What am I missing here?

Dan Grossman

Good thing should != can

Moshe

about these inline images, that is?

Wes

9:00 PM

@rchern what exactly are you doing?

Tim Stone

@rchern You shouldn't have too much to do, because you thoroughly documented all of your source code, right? ;)

drachenstern

@Moshe I was talking about @DanGrossman

rchern

@TimStone Umm. Errr. Uhh. Well.

Tim Stone

Heheh

drachenstern

ok apparently I have no clue what the name of that technique is, because inline image gives a whole other result on goog

Wes

9:02 PM

@drachenstern are you talking about data urls?

Dan Grossman

Data URI scheme

The data URI scheme is a URI scheme that provides a way to include data in-line in web pages as if they were external resources. It tends to be simpler than other inclusion methods, such as MIME with cid or mid URIs. Data URIs are sometimes called Uniform Resource Locators, although they do not actually locate anything remote. The data URI scheme is defined in RFC 2397 of the Internet Engineering Task Force (IETF). The IETF published the data URI specification in 1998 as Proposed Standard on the IETF Standards Track, and hasn't progressed it since. The HTML 4.01 specification refers to th...

drachenstern

mark.koli.ch/2009/07/…

Rejoice rejoice kbd is back

@DanGrossman Data URLs don't work in IE6 and IE7, and are size limited in IE8 and work only for CSS images. Not a good choice (unless you don't need to support IE)

drachenstern

@Rejoicerejoicekbdisback the question was could it be done to consolidate the pages into one download

we were really discussing the theoretically accomplishable feat

rchern

Actually my code is decently documented, I just do a lot of other stuff too.

Wes

9:03 PM

Data URI scheme

The data URI scheme is a URI scheme that provides a way to include data in-line in web pages as if they were external resources. It tends to be simpler than other inclusion methods, such as MIME with cid or mid URIs. Data URIs are sometimes called Uniform Resource Locators, although they do not actually locate anything remote. The data URI scheme is defined in RFC 2397 of the Internet Engineering Task Force (IETF). The IETF published the data URI specification in 1998 as Proposed Standard on the IETF Standards Track, and hasn't progressed it since. The HTML 4.01 specification refers to th...

Rejoice rejoice kbd is back

@drachenstern ah. Well theoretically... And IE has this proprietary HTML + images format I think, I forget the name....

drachenstern

@rchern so overworked and underpaid huh?

Dan Grossman

One HUGE sprite which JavaScript throws on a canvas then slices up into several images

Tim Stone

Ah, well that's a good start then :P

drachenstern

@DanGrossman that's my thought process, and as a DataURI, well... ;)

rchern

9:04 PM

@drachenstern for a couple more weeks. (;

Dan Grossman

Once someone writes OCR for Canvas, you can even embed the text as part of the image! The image becomes the container of the whole page, with a 1-line HTML file to invoke the JavaScript to build it.

Wes

@Rejoicerejoicekbdisback yeah it does. Html archive I think its called but what use is that?

Dan Grossman

I wonder how densely you can do an image-to-text encoding, something like QR codes

Wes

@DanGrossman css spriting isn't appropiate for foreground images.

example. Gravatars in this chat are part of the contents of the page.

Moshe

I've decided that the web is closer to procedural programming than it is to OOP.

Wes

9:07 PM

Damnit my wifes on some video chat. Everything is timeing out including this room.

Tim Stone

scratches head

Moshe

The way we write web code makes web pages into spaghetti code .

@TimStone - Why?

JS is OOP, but to a point.

HTML has tags, but tags are not objects.

Dan Grossman

HTML isn't code, it's markup

drachenstern

@Moshe no it's not

Moshe

@DanGrossman - True.

@drachenstern - JS is not OOP? I think it is to a point.

Wes

9:08 PM

in Let's get philosophical, Nov 27 at 15:10, by Wes

ooh I wonder if you could use client side XSLT to transform from a standard xhtml layout to an epic xhtml layout. No extra elements.

Moshe

Well, i guess it's not any more than HTML or CSS, so perhaps you're right.

@Wes ???

drachenstern

Javascript is a Multi-paradigm: prototype-based, functional,[1] imperative, scripting language

Dan Grossman

new Object()

Moshe

Ah, ok.

good point.

Wes

you should use xml for the semantics. Everything else should be bolted on. Use server side XSLT to xhtml (good for search engines) and use xslt from XHTML to XHTML + non semantic markup (nested divs for example)

Dan Grossman

9:10 PM

object-oriented programming is as much about what you write as what features the language has and what it calls them, no?

Moshe

@DanGrossman - I'm not arguing with that.

drachenstern

@DanGrossman remember Atwood's Law: "You can write FORTRAN in any language"

Moshe

@Wes - That's what the standards people said in 2002. Use XML, thus XHTML was born.

drachenstern

@DanGrossman I think it's about whether objects are first order or second order citizens in the language topography

Moshe

Theoretically, when the web was primarily used for colleges and government agencies, all you needed was markup. Semantics work when you're transmitting meaningful data.

Wes

9:12 PM

Yeah not actually xhtml which btw I mark my pages as being. I mean XML so you have clear content.

Actually I'm rather fed up of pretty sites.

Moshe

As far as the end user goes, a youtube video, while being data on the technical side, is not data!

drachenstern

afk

Moshe

So if I want to make a video site or a photo album or a web application, why should I be forced to use a procedural language to mark up everything as data?

@drachenstern ???

I'm not doing well with the acronyms today. afk?

Wes

AAH timeouts

@Moshe what do you mean its not data?

Moshe

Conceptually, a video isn't data.

Data is an address, a phone number, a name.

A video has attributes that are data, but for a web designer, I can't accurately create xml tags for everything.

Sure everything is represented by data in the back end, but I don't see how "semantics" fits into programming and things like user experience and the like.

Wes

9:16 PM

Ah your talking about binary data. The xml can tell you what the binary data is though.

Moshe

right, but it's not binary data.

Why must the web know the semantics of my page?

iOs doesn't do that, nor deos Java.

Wes

I'll bet I'm timed out again

Moshe

@Wes - log into the router and throttle your wifes connection, or boot her. (or don't)

Wes

@Moche If my router had QOS I'd have trottled my stepson back to 128k by now

Moshe

Just kick him off for a bit.

Cant' do that?

Wes

9:20 PM

I pull out his cable every so often yeah :P

works surprisingly well for a bit.

I could MAC address filter I suppose.

Pinging google.com [74.125.230.80] with 32 bytes of data
Reply from 74.125.230.80: bytes=32 time=608ms TTL=51
Reply from 74.125.230.80: bytes=32 time=718ms TTL=50
Reply from 74.125.230.80: bytes=32 time=600ms TTL=50
Reply from 74.125.230.80: bytes=32 time=314ms TTL=51

Ping statistics for 74.125.230.80:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 314ms, Maximum = 718ms, Average = 560ms

Moshe

ouch

Wes

Yeah owch

okay thats better bdown to 150 ms or so.

drachenstern

@Moshe The web does NOT need to know the semantics of anything, but there is a movement for the semantic web, where we can make things more intelligent based on what is presented

Moshe

@drachenstern - well then perhaps we can have a movement for a semantic desktop too?

drachenstern

A browser can still display text files, and a server can present text files to the browser

Wes

9:27 PM

It only makes sense in the end. The web's moving off just browsers as a display medium.

Moshe

I think it's ridiculous.

drachenstern

@Moshe that's sort of happening

@Wes depends on how you define "the web" ;)

do you mean TCP/IP?

Moshe

Semantics are only needed in academia where everything must be searchable.

drachenstern

Do you mean http?

do you mean html?

Wes

No actually I mean http / html

drachenstern

9:28 PM

@Moshe WRONG, try google?

Moshe

0

Q: Why is the web still interpreted HTML and CSS and JS?

Why is the web still using multiple independent languages to make sites? why not use a single OO language and call it a day? In 2010, when bandwidth is many times more than what it was in 1995, why should we not be compiling web pages or translating them to some intermediate language? HTML and J...

@drachenstern - I understand that that's how it is now, but it shouldn't necessarily be so.

drachenstern

wired.com/magazine/2010/08/ff_webrip/all/1

Moshe

what needs to be searched?

drachenstern

@Moshe why not?

Moshe

Your name?

Privacy invasions, for one.

Clogging servers, IP addresses and bandwidth for another.

drachenstern

9:29 PM

@Moshe you should only be marking things up that need to be marked up ... that's the point of hypertext MARKUP language

Moshe

We don't need to have a copy of everything Cached at google.

drachenstern

@Moshe how will they get content?

Moshe

@drachenstern - It becomes impossible to not markup everything for CSS and JS reasons..

That's my point.

@drachenstern what about building a web app?

Do I parse my code for specific text using javascript?

Wes

wow @Moshe you've the exact opposite opinion to me

Moshe

CSS can't handle that.

drachenstern

9:30 PM

wait wait wait

Moshe

@Wes , what's your opinion?

drachenstern

let's be clear about something

Moshe

@drachenstern I'm here.

@drachenstern - ok

drachenstern

in Java, how do you build an app? how do you define the textboxes? how do you find the data in the files that you want?

Moshe

In text, and it's then compiled.

Wes

9:31 PM

Everything should have a meaning attached behind it. All elements should mean something so the interperation of those elements isn't up to the viewer.

drachenstern

The questions you're asking make sense if you assume that the web is MERELY text, but it's not.

Wes

And most sites should look/work the same

drachenstern

@Wes they do. they have links and input boxes and text and forms and buttons and selects and ... get the idea? theming will NEVER be consistent, look to Milan or Paris for the reasons why not

Moshe

@drachenstern please explain.

@Wes, Most sites do look and work the same, which kills novelty.

Wes

Yeah then it ouly works on teh few browsers you ftested.

drachenstern

9:32 PM

it's not just text, it's hypertextmarkuplanguage and it's ecmascript and it's contentstylespecification

Moshe

@drachenstern - you mean cascadingstylesheets

Wes

Novelty is confusing people spend nearly 100% of their time on other sites.

Moshe

:)

Wes

Being different isn't good its confusing

drachenstern

if you have a website that has only text, then you don't need those things

Moshe

9:33 PM

@Wes - People don't want uniformity. CSS is the proof.

drachenstern

@Moshe perhaps so ;)

Wes

@Moche designers don't want uniformity

Moshe

If people didn't want novelty, they would make b&w webpages.

Wes

users do

drachenstern

but back to my waitwaitwait

Moshe

9:33 PM

im backing to your waitwaitwait

Wes

Saying people don't want peace. Guns are the proof.

drachenstern

how do you tell the Java framework exactly which part of the form to manipulate?

Moshe

@drachenstern - I'm not a java guy, you tell me.

A pointer, probably.

drachenstern

ok, what language do you write in?

Moshe

I've done some Java, barely.

I can do...

drachenstern

9:35 PM

so nominally you write in ?

Moshe

PHP, ActionScript3, Objective-C for iOS..

drachenstern

ok, objc

Moshe

Mostly web and lately iOS.

I've done some VB.NET too

drachenstern

how do you tell the code which elements on screen to manipulate?

Moshe

you hard code the pointers and pass them values.

drachenstern

9:36 PM

so what if you could soft-code* the pointers and still pass them the same?

Moshe

@drachenstern - what if?

drachenstern

soft-code* late binding? merely meaning if you didn't know ahead of time the values to be hard-coded (since hard-code means string literals and the like)

Moshe

Ok

your point is?

drachenstern

so if you could select the elements from your obj-c interface at run-time and grab the elements to work with them

would it change how you work with the elements?

Moshe

quite possibley, so?

drachenstern

9:37 PM

that's all javascript does

Moshe

ok, but your "objects" are html tags.

drachenstern

otherwise the code is written nearly the same (ok, a lot less apis) and it is still compiled to bytecode

so?

Moshe

Incredibly inconvenient.

drachenstern

what are the "objects" in obj-c?

Moshe

And everything is linear.

drachenstern

9:38 PM

what is incredibly inconvienent?

memory is linear

Tim Stone

O_o

drachenstern

hard disks are linear

Moshe

String literals.

drachenstern

cdroms are incredibly linear

Wes

If you had binary objects in there who would set the standard. How would google understand your page. How would browser 2 understand the page.

drachenstern

9:38 PM

everything in javascript is not a string literal, nor is every resource in HTML land

Wes

Would a web page only work on win32

Moshe

Right, but I don't need to care about tht. (The cdroms and hdd)

Wes

have to be recompiled to work on linux, recompiled to work on win64, solaris, etc.

Moshe

@Wes - Tha'ts something the W3C could add to their list of things to waste time on.

drachenstern

@Moshe nor do I but I don't understand what your point is about it's all linear

Wes

9:39 PM

You want to tie the browser down to sending binaries accross the way.

Moshe

@Wes - nope, an intermediate language.

Wes

Its interpreted becasues it allows for multiple interperters.

Moshe

My point is that the markup is silly.

drachenstern

woohoo, IL for the web!

Moshe

So make it interpreted, as is Java.

drachenstern

9:40 PM

my point is that the markup exists in other languages

Wes

@Moshe what would the language be.

Moshe

@Wes, who cares

drachenstern

if I open a form written in VB6, it's still just markup

Moshe

@drachenstern - It's not markup.

drachenstern

it is sir

Wes

9:40 PM

You want to specifiy behaviours not content.

as in applications not repositorys of knowledge right?

Moshe

@Wes, that sounds about right.

@drachenstern - no markup is not code.

drachenstern

WRONG

Moshe

XML is not C

drachenstern

all code is markup, of some sort or another

we're meerly telling the computer what to do given certain objects

Moshe

Okay, I see what you're saying,but ho does that change my approach?

Wes

9:42 PM

well there may be some people doing binary on limitied chipsets. as a hobby but otherwise all code is markup he's right.

Moshe

@drachenstern yes, but HTML needs to be coded in a certain order and it's extremely limiting.

Nick T

So do most programming languages (coded in a certain order)

Moshe

ObjectiveC gives me more freedom in the IDE and in my source files.

Okay, I see I'm having a rough day with this argument. Ugh.

drachenstern

@Moshe here's a snippet of code from some random .frm from before I started here on the network drive:

Moshe

.frm is VB6?

Nick T

9:43 PM

Though languages where you're obliged to write dozens of lines of boilerplate code for every program are dumb.

drachenstern

VERSION 5.00
Begin VB.Form REDACTED_DB_MESSAGE
   Caption         =   "DB Connection...."
   ClientHeight    =   1125
   ClientLeft      =   60
   ClientTop       =   345
   ClientWidth     =   5400
   Icon            =   "REDACTED_DB_MESSAGE.frx":0000
   LinkTopic       =   "Form1"
   ScaleHeight     =   1125
   ScaleWidth      =   5400
   StartUpPosition =   3  'Windows Default
   Begin VB.Label Label1
      Caption         =   "Connecting to SQL Server. Please wait !"
      BeginProperty Font

Wes

@Moshe I'm even against html5. But for the opposite reasons you are. I think we should be moving the applications out of our content documents not enriching them. I am in favour as I've mentioned in this room of an application language that can embed documents not the other way arround.

drachenstern

@Moshe yes, VB6, but my point is, look at the layout and the requirement of certain things to be a certain way

Moshe

@Wes - I kinda agree with that. I hate HTML5 too.

drachenstern

on screen it looks like a window

but it's really just a file with some lines of text that describe how the window should look

Moshe

9:44 PM

@drachenstern - true

Wes

But I want markup to have meaning. otherwise its irrelevant, hard to interpert / compile

Moshe

Maybe it's some affinity I have for iOS.

drachenstern

which is almost identical to HTML

Moshe

I just love the platform.

drachenstern

and this is how forms for windows are designed in general

I can do the same thing with a modern C#.NET project that has a form

Moshe

9:45 PM

Fair enough.

You've convinced me.

drachenstern

so my point is they are all semantic objects in text, that's all, I just don't see where compiling gives us a benefit ...

@Moshe for some reason I don't think so but I'll accept and shut up ;)

Moshe

@drachenstern - you are right about the markup point.

I just don't know if HTML's style is the right one.

drachenstern

well we've outgrown it, hence the desire for HTML5

Moshe

@drachenstern - Ok. btw:

-1

Q: Why is the web still interpreted HTML and CSS and JS?

Why is the web still using multiple independent languages to make sites? why not use a single OO language and call it a day? In 2010, when bandwidth is many times more than what it was in 1995, why should we not be compiling web pages or translating them to some intermediate language? HTML and J...

programming-languages tools html

drachenstern

but as for the separation of CSS from HTML, that sort of thing, go back and look at the example .frm I posted, see where each attribute has all the styling defined within, there's no inheritability there, aside from programmatically

Moshe

9:48 PM

Hence subclassing.

Wes

No its more than that

drachenstern

subclassing meaning? in terms of obj-c?

that's not what I mean ... I mean how can you define all the buttons in VB6 to have the same background color? you either programmatically update them all or you do it by hand. There's no common "class" (to use the CSS term) for buttons to share ... I don't know that there's such a thing in modern C#, although yes, I could do inheritance there.

Wes

btw there is a lot of binary stuff on the web too. MSN, video conferencing. But very few open standards for the binary stuff.

drachenstern

by the nature of it being binary, of course ;)

Moshe

That's what I'm saying, most modern languages offer a superclass for many things.

Use them. ;-)

Wes

9:50 PM

@drachenstern you could sortof do it in vb classic too.

drachenstern

that's why we have common standards like the Joint Pictures Executive Group or Portable Network Graphics or Motion Picture Executive Group etc ;)

@Wes yeah but it wasn't a first class citizen was it?

Moshe

a'ight

drachenstern

in HTML/JS/CSS it is

@Moshe that I would love to see in HTML

Wes

true

Moshe

that is a strong point of CSS.

Subclassing, or just plain classes.

Wes

9:51 PM

People strive to sperate these things to make their lives easier.

Moshe

But you are suggesting HTML5 support subclasses?

Great idea!

drachenstern

it's sad that we have to assign multiple names in CSS too... I'm waiting on CSS to support inner classes or nested classes

Wes

@drachenstern huh?

drachenstern

input.myButtons { #red { color: red; } #blue {color: blue; } background-color:green; }

Wes

XHTML2 sortof had something going for it.

drachenstern

9:52 PM

that's a bad example but it works

Moshe

@Wes XHTML1.1 or higher never worked for me.

@drachenstern it supports it already.

#parentid > #childid

no?

Wes

it doesn't work at least in the real world. no. There was never support built for it.

some ideas are really good though. Hopefully some of these see the HTML5 working draft.

drachenstern

@Moshe not the same thing

I'm not talking about children selectors

Moshe

oh ok

drachenstern

I'm talking about input.myButton.Red, but also having input.YourButton.Red and have them mean two different things...

it's stupid sketchy without an example, but effectively I mean just writing less overall CSS

Wes

9:55 PM

why can't you? you can have multiple classes.

Moshe

ok, I'm out. I finally got some HTML working for the first time in a long time, so I'm less bitter about it.

Gnight folks.

Wes

oh I think you mean for example mybutton.red means red text for example yourbotton.red means red background?

Tim Stone

He's talking about how CSS forces a lot of repetition.

Moshe

Thanks @drachenstern for the time.

drachenstern

@Moshe anytime

Moshe

9:55 PM

@TimStone - oh so you're here.

gnite

:-)

drachenstern

@TimStone yeah but I can't find a good example

G'night.

@Moche good night

@Wes yeah like that

I'd shoot any of my developers if they put in a class called .red

for the record :P

Tim Stone

9:56 PM

For the sake of illustration.. :P

Wes

yeah I know I was trying to lighten it up a bit

Moshe

@wes it's Mo*s*he, btw ;-)

drachenstern

@Wes but why couldn't we if it was that simple, if things nested like that?

Wes

@Moshe sorry. I don't pernounce your names in my head well.

Moshe

@wes -s'all good nite.

Wes

9:59 PM

@drachenstern were you and I talking about the same things there? I'm not sure do you support semantics in markup or not.

drachenstern

I do support semantic markup on things which have semantics

2

I'm not a fan of marking something up just because I can

Transcript for