The Net’s top standards body is getting closer to speeding up XML-based software, a move that could benefit everyone from cell phone carriers to television broadcasters to the military.
Faster XML ahead?
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
WTF is with all the hype about XML lately? It’s not as if it hasn’t already been around for ~30 years in one form or another.
People are like XML-THIS XML-THAT, when there are ALTERNATIVE and BETTER solutions.
“People are like XML-THIS XML-THAT, when there are ALTERNATIVE and BETTER solutions.”
Let me guess? S-Expressions.
Try: EDI.
People have XML on the Brain. The current disease/fad in IT.
XML is text. With tags. That’s it.
representing data using XML is simply too bulky
and
We’ve started using it in all kinds of situations that it wasn’t designed for.
From a bandwidth perspective, it is one of the most wasteful formats ever, and saddles you with a parsing engine that ends up requiring more horsepower than should really be needed.
The only advantage it has is you can sit down and read it as plain text, which of course makes it OH SO secure.
Ah, time to post one of my favourite quotes about XML:
“XML is not the answer. It is not even the question. To paraphrase Jamie Zawinski on regular expressions, “Some people, when confronted with a problem, think “I know, I’ll use XML.” Now they have two problems.””
taken from http://dirtsimple.org/2004/12/python-is-not-java.html
not sure why there is so much discussion on this topic…gzip is stable, well understood, compresses well….webservers already use it
That was a wonderful article you linked to.
It was designed to be a cross platform data swapping format. And now everyone has grabbed the term and tried to morph it into everything. It was great for its original reason. But now its gotten out of hand.
binary XML. which is ridiculous. if you need something concise, DON’T USE XML.
The thing about XML isn’t in the ML part, there are millions of ways to makup some data. The X is the whole point, eXstensible.
With the use of namespaces and uri’s it’s a breeze to add uncoordinated forward and backward compatible changes to a particular language. XHTML+MathML+SVG is a prime example.
xml is just a buzzword. industry hype. just like the semantic web. and like the semantic web, it gets government funding if you sprinkle your grant proposals with these magic words.
binary XML. you’re joking, right?
Well, the XML hype can’t be worse than the UML hype.
As long as there was one only binary format and there was a BSD-licensed library to make a one-to-one translation from binary to standard XML than I don’t really see a problem as it makes the changes below the logic layer of the program.
Unfortunately the error here rests solely in the bad article and synopsis.
The problem is not so much about making XML binary as it is about having to do this in order to *map*, *encapsulate*, and *transport* it over networks that don’t ‘talk XML’ or perhaps not even TCP (i.e. cell, broadcast and mil).
People have been doing this for some time though, using proprietary and non-proprietary mechanisms. Either via translating the data at the transmitter and receiver, or via encapsulating the entire payload as a ‘blob’.
The real good thing of XML is not a matter of technology. The major point is that it is a BROADLY USED standard. That’s why everybody use it : because everbody use it.
That’s the only real purpose and advantage of a standard : being used. Once it is being used, it becomes usefull to use it. Got it ?
Well, with XML I know that at least others can use the file to some degree, more ofc if I provide a DTD!
If I throw a Yaml-File at random persons, I assume those will be deleted, at best.
If you give me a yaml file, I scream “Yeah! He’s got it!” and start coding a Ruby app. 😉
Then you are probably one under a thousand or so to speak.
Well, as I work most with Java or C# however, I would XML just because XML would popup at some places in the framework – and the last thing that is good is having the incosistency of having included two technologies doing exactly the same tasks. In that case technological benefits must be checked agains maintaince and design issues …
Why not use ASN.1? It can be a tagged based binary protocol.
The main difference would be the tags are non-descriptive, which would require a table lookup for interpretation. Plus ASN.1 already exists and has been around since early 80’s.
Because all, XML, YAML and then ASN.1, too, I guess, are for storing data in a common framework. As long it is just your app that just stores and loads data to and from disks, it is no problem what you choose. But when you want to use it for communication, like you do for example with RSS feeds, SOAP, etc., you’ll want to use something that is wide spread, and both Yaml and ASN.1 then seem to fail in this area.