LCA: Addressing the failure of open source [LWN.net]

LCA: Addressing the failure of open source

Posted Jan 18, 2012 7:27 UTC (Wed) by pkern (subscriber, #32883)
In reply to: LCA: Addressing the failure of open source by rsidd
Parent article: LCA: Addressing the failure of open source

Adobe's software is closed source, but the PDF format is sufficiently documented that lots of other programs can read it.

They might claim that they're able to read it. Given the PDF format's insanity it's actually quite hard to do it right and there are still documents that are only readable (or printable) with Adobe's software. Even if the standard might be open.

I don't have a problem with someone sending me a Microsoft Office document anymore, because libreoffice can read it perfectly well most of the time.

My experience tells me that while this is mostly true for the old binary formats, it's not the case for the new XML-based "standards". And given all the discussions I read about this so-called open standard I'm pretty sure it doesn't contain everything you need to know to implement a reader and a writer correctly. Interop is a hard problem. And while it is great to have it (and I really appreciate what the folks of Samba have done), I'd rather like a program that's sufficiently cross-platform and libre with its own file format, than MS dumping yet another file format to implement over the fence. They'll always have an advantage for years while the folks using Linux can't read those files properly, who are then often being forced to use Windows in some way, depending on the company environment.

(Log in to post comments)

It's not so simple...

Posted Jan 18, 2012 13:12 UTC (Wed) by rsidd (subscriber, #2582) [Link]

Anyway, my point was that if we have enough competing programs, they are forced to be compatible with one another. Right now most people care only about Adobe Reader, but people in the free software world perhaps care only about poppler-based readers -- hence bugs on both sides.

Another example is HTML -- once upon a time, there was only one dominant browser, Netscape; then only another dominant browser, IE. There was an open standard and a standards body but the de facto standard was whatever the big guy did. That has changed thanks to competing browsers, especially on mobile platforms. The fact that two of the engines (webkit and gecko) are open source seems, to me, irrelevant, except maybe to the extent that open source developers are more mindful of standards -- but many would argue that Opera is more standards-compliant than Gecko, while webkit, though open-source, has been driven primarily by one company with a very closed-source mindset. The important thing is that there is competition and therefore interoperability is needed. It's the same thing that ensures plug points, screw pitches, pedal layout on cars, etc, are reasonably standardised.

It's not so simple...

Posted Jan 18, 2012 15:12 UTC (Wed) by jwakely (subscriber, #60262) [Link]

> The fact that two of the engines (webkit and gecko) are open source seems, to me, irrelevant

So you don't think that Safari and Chrome would use two separate engines if they hadn't started from a common, open source code base?

And the fact that you say "gecko" not "firefox" is nothing to do with the code being open source? Would you be able to choose from dozens of free browsers (midoria, galeon, epiphany etc.) and know they're render the page well because they use one of the same two engines? Would all those projects be able to rely on well-tested, standard-compliant engines if those engines weren't open source? Wouldn't they all have their own, buggy engines with slightly different behaviour, or just not exist at all?

If the browser market was split into lots of different engines no single one would ever have competed with IE's dominance, and web designers would have continued to write for IE only. The fact that gecko and webkit have a significant share, and MS were forced to make IE more standards compliant, is partly due their use in multiple different browsers, which is possible because they are open source.

It's not so simple...

Posted Jan 18, 2012 15:28 UTC (Wed) by rsidd (subscriber, #2582) [Link]

Of the gecko-based browsers, only firefox counts, and of the webkit browsers, only safari and chrome count - as far as webpage designers are concerned. Even without chrome, my argument stands. Apple forked webkit from khtml, which had an insignificant presence earlier; if Apple had written their own engine from scratch, it would still have benefited standards compliance and browser neutrality among webpages.

Your Webkit history is wrong

Posted Jan 18, 2012 16:11 UTC (Wed) by robla (subscriber, #424) [Link]

Webkit started out life as KHTML. While Apple now has all the resources it could want, the Apple that started Webkit would never had had the resources to start a credible browser initiative from scratch. Open source mattered a great deal then, and it still does.

Apple, Google and many others collaborate on Webkit now. It's a fantastic example of why open source matters.

Your Webkit history is wrong

Posted Jan 18, 2012 16:56 UTC (Wed) by rsidd (subscriber, #2582) [Link]

I know (and said in another comment) that Apple forked webkit from khtml. They also took it much further, very fast. Konqueror with khtml was just about usable, and yes I used kde at that time. I don't see a reason to doubt that the company that developed OS X and all its bundled apps could have written a browser from scratch. (Yes, I know the non-GUI core of OS X is open source too - and I used FreeBSD at the time.) Google picked up webkit much later, and seems to have focused on javascript since (which is not shared with safari) - I don't know whether they have a significant contribution to webkit itself.