I've heard this comment about OpenXML (the xml format of the office documents) before, and i'm a bit on the fence about it.
It's of course indeed ridiculously complex, but so is office. Microsoft both adds a shit ton of functionality to their documents, and keeps an impressive amount of backwards compatibility.
In the past i heard complaints about part of the OpenXML spec that also allows older binary data in there for backwards compatibility reasons, which of course means for OSS implementations that they don't just have to implement this spec, but also the older spec that came before to be truly compatible with everything a modern office version can open.
But on the other hand, if i look at it from the side of Microsoft, they opened up their format, they've got a gazillion functionalities, should they remove functionality to appease the open source developers? If so which? Should they stop being backwards compatible with documents of decades ago to appease the open source developers? If so how long should they support? Are you going to tell their customers?
Office is an immense program with an immense amount of legacy features, backwards compatibility, ....
It's incredibly complex by nature. And might they have made the format more complex to dissuade competition? Could be. However, in this instance Occam's razor pushes me more to "write a huge program over a timespan of many decades, with thousands upon thousands of programmers working on it, and you'll indeed most likely end up with something very complex...."
I would agree, except that every piece of it is significantly more complex than it needs to be. ODF is considerably simpler in part because it makes use of other pre-existing standards for things like dates and times. OOXML redefines so many of those things, and in many cases Microsoft Office's implementation isn't actually compatible with their own standard.
Do you have more concrete examples? I'm reasonably familiar with OpenXML, and seeing the date issues in microsoft systems (Excel having the same bug that considers 1900 a leap year, to stay compatible with Lotus Notes), i can imagine them redefining everything just to be in full control ^^'...
Integer storage in spreadsheets... There are a ridiculous number of ways to store any integer, and I don't just mean because you could theoretically store 1 and 00000001 and they'd be interpreted as the same thing.
I've heard this comment about OpenXML (the xml format of the office documents) before, and i'm a bit on the fence about it.
It's of course indeed ridiculously complex, but so is office. Microsoft both adds a shit ton of functionality to their documents, and keeps an impressive amount of backwards compatibility.
In the past i heard complaints about part of the OpenXML spec that also allows older binary data in there for backwards compatibility reasons, which of course means for OSS implementations that they don't just have to implement this spec, but also the older spec that came before to be truly compatible with everything a modern office version can open.
But on the other hand, if i look at it from the side of Microsoft, they opened up their format, they've got a gazillion functionalities, should they remove functionality to appease the open source developers? If so which? Should they stop being backwards compatible with documents of decades ago to appease the open source developers? If so how long should they support? Are you going to tell their customers?
Office is an immense program with an immense amount of legacy features, backwards compatibility, ....
It's incredibly complex by nature. And might they have made the format more complex to dissuade competition? Could be. However, in this instance Occam's razor pushes me more to "write a huge program over a timespan of many decades, with thousands upon thousands of programmers working on it, and you'll indeed most likely end up with something very complex...."
I would agree, except that every piece of it is significantly more complex than it needs to be. ODF is considerably simpler in part because it makes use of other pre-existing standards for things like dates and times. OOXML redefines so many of those things, and in many cases Microsoft Office's implementation isn't actually compatible with their own standard.
Do you have more concrete examples? I'm reasonably familiar with OpenXML, and seeing the date issues in microsoft systems (Excel having the same bug that considers 1900 a leap year, to stay compatible with Lotus Notes), i can imagine them redefining everything just to be in full control ^^'...
Integer storage in spreadsheets... There are a ridiculous number of ways to store any integer, and I don't just mean because you could theoretically store
1
and00000001
and they'd be interpreted as the same thing.