Why is Adobe Acrobat such a bloated pig?

[Tellos Athenaios]

PDF is Adobe's tool to showcase lots and lots of typography and document formatting stuff. Acrobat Reader in turn is Adobe's tool for pushing PDF, which in turn should drive sales of their not-so-cheap designer/publisher tools, which is where Adobe makes its money.

So far so simple if sinful. But here's the kicker: PDF itself could conceivably be the mother of all that is bloated, convoluted, complex, insecure and otherwise exquisitely finger-nail-pulling painful to work with. Adobe is of course no stranger to abuse as an artform (take that, Microsoft!). You may know about "psd is not my favourite file format".

[Lemur]

PDF itself could conceivably be the mother of all that is bloated, convoluted, complex, insecure and otherwise exquisitely finger-nail-pulling painful to work with.

And yet PDF is supposed to be an open format, which would suggest that the container/format itself is not to blame. I like having a standard print/form/readable thingy that you can send people with a reasonable expectation that what they see will be similar to what you sent. As opposed to, say, the multitudinous, conflicting, undocumented versions of DOC.

Very much pro-PDF. I'm just experiencing a surge of Adobe Reader rage.

[Tellos Athenaios]

Don't equate being an “open” format with being a good format (for automatic processing, i.e. rendering). Examplum gratum: C++ source code.

The point is not that PDF is evil, the point is that it is complex. C++ source code, to continue the example, is in fact so complex that it is not decidable what a given statement means because of all the what-ifs involved. (And in fact implementing C++ code, that is, translating it into something which can be executed by a machine is even more complex -- so complex that the spec doesn't really know what it should behave like either. That's what “undefined” means in the context of "the behaviour of i=i++; is undefined".) Complexity means that there's lots of different what-ifs to consider, which balloons your code and slows it down. For instance, did you know that PDF is supposed to be able to contain JavaScript? Whoever thought that was a good idea was either secretly bankrolled by Russian mobsters or just really really looking for security trouble.

Most PDF readers simply deal with this sort of complexity/security issue by not supporting features or culling the deprecated ones. Or just relying on someone else to have gone through the pain & horror & suffering before and using whatever code they came up with. Adobe is however in wedlock (literally locked) into supporting every intricacy and every stupidity which ever made it into the PDF spec and can't afford to just pick the good/sensible bits. And of course Acrobat Reader probably carries around all the code necessary for writing PDF as well.

EDIT: The reboot thing is actually just stupidity on Adobe's end. That or pure laziness, Acrobat is not a critical system component so it should not require reboots. It's not that hard to check if acrobat is running and if so refuse to install until it is closed. (Presumably what they are worried about is ABI breakage, that is if code is replaced and then reloaded strange bugs will manifest themselves if the app is assuming a version 1 of the ABI and you've just moved to version 2.)

[Lemur]

Don't equate being an “open” format with being a good format (for automatic processing, i.e. rendering).

It's true, I do tend to suffer from open bias. I expect better things from published, documented formats. As you lucidly point out, open is no panacea. But hey, at least I know my prejudices and acknowledge them.