Keith Devens .com

I agree XML is bloated, but many of the bandwidth and bloat considerations can be dealt with via some form of compression, at least for the transport phase of things.

My major gripe with XML is not so much technical as it is "technocratic", by this I mean the whole idea of XML as I understood it back in the hey dey 1997 was that it will make it easy for web-sites (or other IT areas) to dessimate data to agents or bots, making it trivial to interchange data. And to some degree of success things like RSS newsfeed fulfill that promise. But aside from that, how many things do you really rely on XML in day to day coding, for me, not many. Flat-files (CSV) are still prevelant for most data exchange, after it gets imported into SQL I'm done.

But aside from that , what most folks don't realize is that most web-sites, companies, IT depts, generally, don't want to make it easy to share data. That's why today people are more intersted in builing html-scrapers, rather than just parsing XML... Outside the open source community , valuable or semi-valuable data is guarded and treasured and not ment to be shared... defeating the whole dream of XML.

Users are discovering that XML is taking huge amounts of bandwidth and is generally very slow compared to other methods.

Huh? If he's referring to the RSS problem, it's got nothing to do with the fact that it's XML and everything to do with the fact that visitors are "sticky" and they usually update by polling.

If he's referring to the RSS problem

He didn't mention that, and I'm pretty sure that's not what he was talking about. Based on the links he gave, he referenced the effort to create a "binary XML". I don't think the binary XML camp is motivated by RSS at all either. They want fast data-interchange between applications and slow devices such as cell-phones. I suppose the XML Binary Characterization Use Cases document would be a good place to look for their motivations. I refrain from taking an opinion on whether binary XML is a good idea -- I'm merely pointing out that Sassenrath's comments had nothing to do with RSS.

I agree that the XML syntax takes up pretty much space when compared with how much space it COULD take up.

But, on the other hand, as Tony says, XML can be compressed, either using generic compression algorithms, or specialized XML compression schemes. And when deploying these algorithms you get a data stream that takes up no more space than an equal optimized binary stream would.

So if looking purely at the space that an XML stream takes up, I don't think there's any need to criticize it. But, of course, there IS some overhead involved with doing the compression and decompression, plus I'm sure that the parsing of an XML stream takes SOME more time than it would take to parse a specially made binary stream.

But I think, as long as CPU and parsing power isn't any bottle neck in your project XML suits it's purpose just fine.