Robert Cailliau Talks With WikiNews

David Gerard writes "Wikipedia's citizen journalism sister site, Wikinews, has a long and interesting interview with Robert Cailliau, who worked with Tim Berners-Lee to create the World Wide Web. 'I also remember a big resistance against PostScript, but what do we see now? PDF everywhere. Fortunately PDF is an open standard and it's fairly elegant, but it could have turned out much worse. SVG did not make it. Tim, who had a longer experience with the internet world, convinced me that the web could only survive if all the code was freely available for everyone who wanted to tinker with it. In 1992-1993 I then worked patiently for some 6 months with CERN's Legal Service to draft a document that put the source code into the public domain. This also implied working to convince the management, up to the Directors, of the need to do so. The result was the document signed on 30 April 1993 that gave the WWW technology to the world.'"

US vs THEM

[Anonymous Coward]

"The result was the document signed on 30 April 1993 that gave the WWW technology to the world."

A telling difference between Europe and the US. If it had been an American with this idea, the line would have read "The result was the document signed on 30 April 1993 that made me a multi-billionaire."

Gopher was not free - it failed

[Shirotae]

Gopher was around before WWW but to run a Gopher server you had to do some sort of deal with some US University or something of the kind. Where I worked, it was just too much effort to try to get our management to negotiate the necessary agreement to put up a Gopher server so after a few brief internal experiments we dumped Gopher. We could just download an run a web server on the same host that was serving FTP and provide more convenient access to the data that was already public.

If getting started on the

Re:

[slapout]

Yeah, it's not like the US invented the internet or the protocols and just give them away or anything.

SVG did not make it?

[Max Romantschuk]

I beg to differ. Maybe if failing is the same as "Not ubiquitous on the web", but I find myself using SVG more and more. My vector work in Inkscape is saved in SVG. I've created dynamically generated SVG and rendered it to static images using Batik, to automatically generate hundreds of heading images for websites. Firefox now supports basic SVG. I wouldn't call it a failure as much as slow adoption...

Re:

[Spad]

Until all the major browsers ship with SVG support built in and enabled by default, SVG will not be a "success". It will instead be relegated to VRML territory - useful but rarely used outside of certain communities.

Re:

[nyctopterus]

I think SVG will make it, eventually, but it lost out to flash in a big way.

The reason, I think, was editors. If you make something in Flash (the editor) it will play in the flash plugin (version permitting of course). But on the SVG side, you had a bunch of things which could export to SVG, but would let you do a bunch of stuff that couldn't be exported. Drawing something for SVG has always been a hit and miss affair -- because it's been so easy to add things that aren't supported by SVG. Having to think t

Re:

[morgan_greywolf]

Inkscape doesn't have floating palettes? That's news to me [inkscape.org]. And it doesn't have a Mac OS X version [modevia.com]? Oh, you probably mean a package that doesn't require X. I don't see what the big deal is about supporting native Cocoa widgets anyway. It's not like packages like Photoshop and Illustrator haven't veered away from the system standard widgets for years anyway (those palettes are Adobe's proprietary widgets, not native Mac OS widgets). If it's really that big a deal, then why doesn't someone take the sour

Re:

[nyctopterus]

Sigh, this is such a problem with many open source projects - they get 95% there, and then say who needs that fancy nice shit on top, it basically works. If you want it, write it yourself! Well, newsflash: artists aren't programmers! I think that is why the open source creative apps have seen little success in comparison to system/server applications.

The big deal about X11 is that it's as slow as hell (Inkscape takes 3-5 minutes to boot on my system, what's that about?), the fonts look like shite (might not

Re:

[morgan_greywolf]

The big deal about X11 is that it's as slow as hell (Inkscape takes 3-5 minutes to boot on my system,

What kind of system? Older systems might take a bit. Plus, is that with or without X11 started? If it's without, that's kinda unfair.Inkscape loads up nice and fast on a Windows or a Linux box. Faster than Illustrator, Corel Draw, Xara, or any other graphic design/layout app I've used.

Inkscape has SOME floating palettes, yes. But it needs floating palettes for colour, gradients and layers. It would also be nice if they didn't take up half the screen.

Inkscape has floating palettes for all of the above. What version are you on?

If none that stuff bothers you, I can guarantee you've never worked on images/design professionally, nor even as a serious hobby. This stuff needs to be fixed, or Inkscape is heading to GIMPsville.

Surprise! I once made my living for more than 5 years as a professional graphic designer.

Re:

[nyctopterus]

With X11 open, on an older G4 system it's true, but Illustrator and Photoshop take around a minute to boot. I don't think this is a serious problem with Inkscape per se, but it seems to be fairly common among X11 apps, which goes to my point about X11 sucking. Also, the keyboard shortcuts being Linux/Windows style (ie. ctl+ rather than command+ is a serious annoyance on a Mac).

By "floating palette", I mean the OS X-style utility windows, which don't have to be activated. Inkscape has windows for such things

Re:SVG did not make it?

[iamdrscience]

I like SVG, but I really don't see why so many people have been taking issue with this guy's statement that it has been a failure. First of all, it's pretty clear that he was talking about SVG failing as a web graphics format and in this regard I think he's completely right, I can't think of any page I've ever seen that embedded SVG images in them other than SVG example sites. It doesn't even seem to be gaining much steam now that more browsers support it, Firefox and Opera support natively and although IE doesn't, Adobe includes an SVG plug-in for IE with installations of Acrobat Reader (or at least they did, I haven't checked lately). The only time I regularly run into SVGs on the web is on sites like Wikipedia or sometimes F/OSS project sites, but even there, they're never embedded in the page, they're have a rasterized version of the SVG and then link to the SVG file as a "source" for it.

Outside of the web, I agree, it would be unfair to call SVG a failure, but that said, it hasn't been a runaway success either. SVG has been successful enough that people use it and it's generally well supported by most vector drawing applications, but most people don't work using SVGs, they use whatever format is native to their application (Inkscape users being an obvious exception because its native format effectively is SVG). Also, while SVG has gained acceptance as a platform/application agnostic way to send vector artwork to other people, it's still less popular than other formats like EPS (or even PDFs nowadays).

Re:

[Mad Bad Rabbit]

[SVG] doesn't even seem to be gaining much steam now that more browsers support it, Firefox and Opera support natively and although IE doesn't, Adobe includes an SVG plug-in for IE with installations of Acrobat Reader (or at least they did, I haven't checked lately).

Adobe has announced they will drop support for the plug-in on January 1st 2008. And I suspect that by then, Firefox still won't provide any controls to pan or zoom embedded SVG images (which leaves it useless for large diagrams, maps, etc.)

Script?

[tepples]

And I suspect that by then, Firefox still won't provide any controls to pan or zoom embedded SVG images (which leaves it useless for large diagrams, maps, etc.)

Isn't that what script is for? Or is there a deficiency of Firefox's script engine that keeps script from doing this?

Re:

[Hatta]

You could have made the same argument a couple years ago for PNG.

Re:

[iamdrscience]

PNGs are a whole different ball of wax, but I still don't think PNGs really make sense for the web today. The reason being that IE6 doesn't support transparencies in PNGs right and PNGs without transparency offer no advantage over JPEGs (besides being a free standard, if that's your thing) and they're significantly larger. Sure, I use PNGs on my personal sites where I can afford to say "fuck you, IE6 users, I want full color and transparency and if you don't like the way it looks, upgrade your shitty browse

One word on PNGs

[Aleksej]

Screenshots.

Re:

[tepples]

The reason being that IE6 doesn't support transparencies in PNGs right and PNGs without transparency offer no advantage over JPEGs

No advantage? JPEG introduces blocky truncation artifacts, which are especially noticeable at edges of flat areas. PNG does not.

Re:

[jandrese]

Those artifacts really only show up when you have the compression turned up too high or you're trying to encode text or something like that. If you turn the compression up to the point where you don't see blocks, you'll still be using fewer bits than the PNG would have. The only general exception to this rule is a screenshot where you have plain 1 pixel thick-lines text. There you will see JPEG artifacting at all but the most insane bitrates.

Re:

[Kadin2048]

Those artifacts really only show up when you have the compression turned up too high or you're trying to encode text

That last bit is why PNG gets used -- if you have a graphic that contains text, you don't want to use JPEG. Hence, it gets used a lot for small graphics that contain text, screenshots, and the like.

JPEG is for photos; PNG is graphics. You don't want to compress graphics that have text in them with JPEG, and it's wasteful to compress big photos with PNG. (Unless you need a lossless format in which case PNG can sometimes be used as an alternative to the many mutually-incompatible flavors of TIFF.) But becaus

Re:

[Hatta]

PNG is an alternative to GIF, not JPEG.

damn, i miss BBS and Fidonet

[randuev]

I miss those times... Bulletin Board software, messages from strangers, file areas, +++ATH0, first multiplayer games...
ps. i wonder how fast would WWW catch on if it was invented today. threat of national security? ;)

SVG failed?

[nagora]

Errr. Wrong, basically.

TWW

I now hate Tim Berners-Lee.

[iamdrscience]

On the technical side it was not always the best of understanding between me and the team. For example, I was convinced that we needed to build-in a programming language, but the developers, Tim first, were very much opposed. It had to remain completely declarative. Maybe, but the net result is that the programming-vacuum filled itself with the most horrible kluge in the history of computing: Javascript.

Amen.

I can't help but think how much further along web applications would be if there were a programming language built-in from the start.

SVG didn't make it?

[saforrest]

Yes, clearly the fact that SVG wasn't used in the manner foreseen indicates it has failed utterly.

I never randomly stumble upon SVGs while browsing the web. Yes, never [wikimedia.org].

Re:

[argent]

First time I'd ever seen an SVG file was when I clicked your link.

First time that you know of, anyway.

Re:

[WillAdams]

Given that there aren't any .svg files on that page, only links to them, and uses .png previews instead pretty much proves the point of SVG not having made it as a main-stream in situ web graphics format, no?

Is there a browser (other than Amaya) which will render .svg files in-place on a .html page?

William

Re:

[dkf]

Is there a browser (other than Amaya) which will render .svg files in-place on a .html page?

Firefox 2.0.0.6 certainly does it.

Re:

[WillAdams]

Please point me to a page which does this.

I just made a test.svg file using Adobe Illustrator CS3, embedding the Minion Pro italic glyphs which I used for ``This is a test'' (w/ an st ligature) for the last two letters), then tried to make a .html file which would display it in-line.

Adobe DreamWeaver wouldn't let me select or drag-drop the .svg file, so I hand-edited it to reference ``test.svg'' --- the graphic didn't display when I loaded the .html page into FireFox 2.0.0.6

Viewing the file directly resulte

Re:

[Tangent128]

Firefox doesn't support SVG in an <img> tag (I don't think anybody else does yet either), but SVG works fine in <embed> and <object> tags.

Re:

[WillAdams]

OIC.

Thanks!

I managed to get it working w/ the embed tag, but couldn't specify the size (I'd really like to be able to do so using a relative em measure). But as the AC noted, there's no font support yet, so the utility is rather limited.

William

Re:

[TheRaven64]

Is there a browser (other than Amaya) which will render .svg files in-place on a .html page?

Gecko has had some SVG support for at least five years, I think it's now enabled by default. Safari 3 supports SVG, as does Opera. IE does if you use the Adobe plugin. The problem with the plugin is that the SVG image is not exposed to the DOM, which eliminates a lot of the advantages of SVG (i.e. you can't modify it with JavaScript).

No one fully implements SVG, but the browsers mentioned above have fairly good support for the subset of SVG aimed at mobile devices.

Re:

[WillAdams]

Safari 3 displays a ``test.svg'' file made w/ Adobe Illustrator CS3 and embedded Minion Pro Italic glyphs using Times. It also won't display a .svg in-line as a graphic in a .html file (instead one gets the missing graphic icon) --- I'm using Safari 3.0.3 --- do you have a link to a page which does use a linked .svg graphic in-line w/ text in a .html page?

Opera also failed to display the embedded Minion Pro Italic glyphs (used Times) and didn't show the ``test.svg'' file in-line in the .html file.

To re-stat

Re:

[TheRaven64]

I just went to the page you linked in Safari 3.0.3. There are two graphics. One is a missing plugin icon. This is the one indicated by this code:

<object id="AdobeSVG" classid="clsid:78156a80-c6a1-4bbf-8e6a-3cd390eeb4e 2"> </object>

I wouldn't expect this to work, since identifying objects by a COM clsid is, as far as I know, IE-only. While this is, technically, a piece of valid HTML, the clsid URI namespace is very Windows-specific. To the right of this, was a red circle with a black b

Re:

[saforrest]

Given that there aren't any .svg files on that page, only links to them, and uses .png previews instead pretty much proves the point of SVG not having made it as a main-stream in situ web graphics format, no?

I didn't realize that MediaWiki used png previews: I'll concede that pretty severely undermines my example.

PDF is irrelevant to the web.

[argent]

Yes, you see PDF everywhere, but in virtually every case it's just another downloaded document format. Yes, you *can* view PDF in your browser, but you don't actually get any particular benefit from doing it that way no matter what Adobe seems to think: the fraction of PDFs that contain hyperlinks is negligible, and Adobe's embedded reader is balky and unreliable by comparison with the standalone one. And PDFs are inherently harder to read... the print-quality rendering and page orientation means that the t

Re:

[MichaelSmith]

As an ad for PDF it showed exactly why PDF isn't an appropriate web technology

I just can't believe that postscript gets used for graphical user interfaces.

Re:

[WillAdams]

It does.

Solaris still comes w/ Display PostScript as an option, and it worked quite even more nicely in NeXTstep, and Display Ghostscript was developed for use in GNUstep.

While Quartz (neé Display PDF, though it adds lots of other things) is nice, I still miss the programmability of Display PostScript and its interactive use in programs like Altsys Virtuoso, which would allow one to program fills and strokes for objects in PostScript.

William

Re:

[TheRaven64]

Solaris still comes w/ Display PostScript as an option, and it worked quite even more nicely in NeXTstep, and Display Ghostscript was developed for use in GNUstep.

DPS in Solaris is Sun's second attempt at using PostScript for GUIs, the first being NeWS. NeWS ran entire user interface widgets in the PostScript interpreter, which made it great for remote display: clicking on a button would give immediate feedback, as the view object was run on the client side, and asynchronously pass an event to the back-end.

One big reason for its failure was that it was hard to write UI components in one language and the rest of the code. It's a shame James Gosling didn't learn

Re:

[nuzak]

It's a shame James Gosling didn't learn from this when he went on to write Java, which could have been NeWS-done-right if RMI had been used by default for communicating with view objects.

Java was already slow out of the gate due to throwing away a couple decades of VM research in its initial version, with the only advancement in the state of the art being the bytecode verifier. I shudder to think of what adding RMI on top of AWT's existing Lack Of Zip would have done.

Anyway, the promise of DPS was that yo

Re:

[dwater]

The older versions of SGI IRIX window manager, 4Sight, used postscript too - Sun's NeWS. Is that the same Sun stuff you're talking about?

I remember the icons in the newer 4Dwm were all 3D too - didn't matter what size you made them, they always looked crisp :) - or was that 4Sight too...so long ago :( It was all good though, and I still prefer it to what is available today.

Display Postscript is a perfect example

[argent]

DPS is a perfect example of what I was referring to, and NeWS is an even better on. Postscript integrated with web pages would have given us the capabilities AJAX provides years earlier, without the problems inherent in the way PDF is designed. PDF barely touches the surface of the capabilities of Postscript, even DPS ignores most of what it can do. Using Postscript as an interchange format or layout created by another program is like using HTML to display images by building tables of one-pixel cells set to

Re:

[rvw]

I know Adobe wanted to push PDF as an alternative to HTML at some point. It didn't happen, and we should be glad it didn't. PDF is for print, and for that purpose it's a good enough format. I prefer it over Word. And HTML with special CSS for print purposes just isn't good enough. So it definitely is relevant to the web today, as a download format for printing content.

Re:

[tryfan]

> PDF is not really relevant to the web today, except as a
> shining example of how not to create content.

I think that depends on the creator - when it comes to long texts that needs a special layout, PDF can work great.
12 - 13 years ago, though, it was a disaster - the slow connections back then made it extremely inconvenient to use PDF documents.
I doubt that we'll ever get past the problem of content vs. looks (which was in fact, one of the basic questions that the Web was meant to solve).

Re:

[xaxa]

The most appropriate example I can think of is when I find a manual for some (usually) free software to download:
- HTML formatted, all one one page
- HTML formatted, one page per section
- PDF
- Plain text
I almost always click one of the HTML versions. If I want to use the inline search, I use the HTML on one page one, otherwise the HTML sectioned one. I don't think I'd ever use the PDF (I never print this kind of stuff).

Most of what's in PDF doesn't need a special layout anyway...

PDF is a decent document distribution format

[argent]

I think that depends on the creator - when it comes to long texts that needs a special layout, PDF can work great.

That's true, there's a lot of material that is not well adapted to HTML. At the moment that means creating a print-quality document and distributing it over the web, and PDF is a decent format for that.

In the context of "being a web page", though, PDF fails badly, and the vast majority of documents distributed in PDF (most of which are not academic papers, they're basically advertising) should b

Re:

[TheRaven64]

the fraction of PDFs that contain hyperlinks is negligible

Most PDFs I read contain hyperlinks. Most academic papers (in Computer Science, at least) are typeset using LaTeX, which generates PDFs natively. If you use the hyperref package then it will use hyperlinks within the document for bibliography and other links (e.g. figures and tables), and lets you create external links easily. Most things I write in LaTeX have hyperlinks to external resources in them.

Re:

[argent]

The fraction of PDFs that are academic papers in computer science are negligible. :)

Most are advertising fliers, real or disguised. :(

Re:

[Watson Ladd]

MathML is very kludge compared to latex. Tell me when MathML becomes Turing complete.

"Windows does not use windows"

[BlueStraggler]

My favourite quote from TFA (in reference to the importance or not of tabbed browsing):

Windows does not use windows

Indeed. I've always thought it should have been named "Screens".

Fascinating interview

[Sinister Stairs]

This is a very fascinating interview -- RTFA, the current discussions here do it disservice.

next-previous Site Navigation still in SeaMonkey

[spage]

Something I do miss are the "next-previous" functions of the NeXT browser. Current browsers only permit you to follow a link and then to run back and forth over the path you took ("back" and "forward"). The NeXT browser had the additional function of following the next link of the previous page ("next"). That allowed me to make a page which was a list of pages to be looked at and then to walk that "path" with a click per page.

I think he's talking about