Old MS Office file formats were never engineered to be future proof in any way, and they're a huge headache for Microsoft, and they do indeed behave oddly in new software, and Microsoft cannot wait to get rid of them, and nobody in their right mind can blame them, honestly.
I'd like to expand on this. The Microsoft Word team is the only group of people that ever had any say over what goes in a Word document. In the old days, a Word document was nothing more than the serialized version of what Word had in memory. Photoshop files are the same way. As soon as you're not the only one with a say over the file format, you turn it into a standard. And you step up to make it a good standard because otherwise nobody's going to use it.
To keep going. Any standard that is not open will not open itself for multiple implementations (see old MS standards with word xml format). The problem of course is that if I cannot implement your standard, you are still the only consumer. So really:
Standards with producers/consumers from different independent teams from different organizations with different goals are awesome.
"Open standards /rock/." is a bit overblown, don't you think? There are still very real problems with open (read: sort of "open") standards. In the end we all just want something that works. There are indeed a great deal of proprietary solutions which satisfy this criteria, and continue to do so for many years.
We're much better off with formats that were documented with inter compatibility in mind in the first place. Historically, it's been quite difficult to build products that work with MSO formats as MS not only did not help meaningfully but actively seeks to hinder others from doing so - plus the format is buggy and odd. Read what Wikipedia has to say on the history of the .doc format:
>Some specifications for Microsoft Office 97 binary file formats were published in 1997 under a restrictive license, but these specifications were removed from online download in 1999. Specifications of later versions of Microsoft Office binary file formats were not publicly available. The DOC format specification was available from Microsoft on request since 2006 under restrictive RAND-Z terms until February 2008. Sun Microsystems and OpenOffice.org reverse engineered the file format. Microsoft released a .DOC format specification under the Microsoft Open Specification Promise. However, this specification does not describe all of the features used by DOC format and reverse engineered work remains necessary.
I don't disagree. I'm just sensitive to the "open standards /rock/" thing. It glosses over a lot of nuance. What I'm trying to suggest is that something being an "open standard" does not necessarily make it better. Historically, there have been a number of open standards which have been failures for technical reasons, political reasons (within the community behind the standard), no good reason, etc.
Yeah, you point out the winners, think back 15 years ago. How many competitors were there just in the realm of word processing? Every format but the eventual winner (which couldn't be determined at the time) is now completely useless, whereas an abundance of old dud open image formats can still be opened (which is nice because some are really easy to programmatically generate)
I don't think that's true. I can still open WordPerfect docs from 15 years ago, with less faff than it takes for, say, VRML (which was an open standard).
He's simply saying that backwards compatibility and open standards are good things in that they don't break old sites/docs/etc...
You could have just as easily said the same thing and ended with "So I don't really see your point about backwards compatibility in this case".
Backwards compatibility is often the forced result of open standards because when multiple vendors adopt those standards it is in their best interest to not make breaking changes.
This in turn is(obviously) a good thing in that things designed for those standards will remain functional indefinitely.
A .doc file created 5 minutes ago in Word can't be assumed to open correctly in any other version of Word without testing.
I used to work at a research institute and the people who used Word always spent more time getting their papers to look as nice as the LaTeX users, particularly if you crossed platforms or had change tracking enabled (hello undeletable footnote in the middle of the page…). The barrier to entry was higher but if you needed reliably consistent formatting there wasn't really a better option, with the exception of HTML5 + CSS + your favorite MathML shim if you're submitting to a really web-centric journal.
> emacs outshines all other editing software in approximately the same way that the noonday sun does the stars. It is not just bigger and brighter; it simply makes everything else vanish [1]
I think if you replace "emacs" with "TeX" and "editing software" with "document formatting system", the quote holds just as well.
PDF was not an open standard until 2008. Just saying. It could have been abandoned by Acrobat/Adobe while it was still proprietary and you would have been left with unreadable files in 2013. That's not the first time these things happen either.
HTML is way more solid in that aspect: as long as you have a browser you should be able to open the document on any browser, any platform and get a very similar access to contents. There's nothing much comparable here... your "doc" example does not make much sense either on platforms where official Microsoft software is not available (Unix, anyone?)
It was not an open standard, but it was nonetheless rigorously documented. If you wanted to modify a PDF (as I did in 2006) it was fairly simple to create an editor in the language of your choice, using only those documents as a guide.
You're lucky you only have to read PDF or DOC files (or as somebody said: "the winners").
These last two months, I've manually parsed old databases[0], some dating back 2001. Locked formats with unknown specs, schemas embedded in applications (good luck digging that out with IDA), data storing that makes no sense, losing your hair out,...
[0] Drop me a mail if you want to reverse-engineer an HyperfileSQL DB from around Y2K, I can help on that if you pay me with Red Bull.
Same for a PDF file created 15 years ago or most other formats. So I don't really see your point about open standards in this case.