Drew McCormack

@drewmccormack So why invent a new partial loading document format when iCloud already has built-in support for one?

@mjtsai You mean Core Data? I think that's pretty clear: the iWork team don't trust it. And maybe the codebase isn't suited.

@drewmccormack Sir, I beg to differ: "There is no option to load an XML document partially". Actually, there is @martinwinter @mjtsai

@avbelow For example: any SAX parser (except NSXMLParser), lazy DOM parsers, MAX, to name a few (partial XML parsers)@drewmccormack @mjtsai

@mpweiher @avbelow @mjtsai Sure, you can partially load, but not much good unless you can save.

@drewmccormack @mjtsai @mpweiher That does not seem to be much more work than (un-)zipping a file on the go, #methinks

@avbelow @mjtsai @mpweiher To save, you will have to load the whole file. Defeats the purpose of partial loading.

@drewmccormack Actually, no: there's a (huge) difference between materializing the objs and pumping the raw XML around. @avbelow @mjtsai

@mpweiher You mean, you read the XML, then write it straight out again to a new file. Still a hit. @avbelow @mjtsai

@drewmccormack PSPDFKit includes a streaming XML parser for XFDF that might contain GBs of embedded sound/image data. libxml to the rescue.

@steipete I'm guessing you are extracting data. That is a bit different to a document format that needs to be in memory.

@drewmccormack Wouldn’t be hard to store the byte address instead - but binary probably is more efficient.

@drewmccormack @mjtsai did many people actually make use of editing the raw XML?

@ObjColumnist @drewmccormack Probably not. I more see the XML as a safety net in case the format is abandoned. I used the PDFs every day.

@mjtsai @drewmccormack good point, but I can't see iWork being suddenly abandoned any time soon.

@ObjColumnist @drewmccormack iWork ’09 is no longer for sale, and the new version can’t read all the features of the old file formats.

@mjtsai @drewmccormack ah so when you upgrade you might lose some things?