In an industry characterized by cold logic and technology, the contributions of real people often get lost. Publicly, great things are seen to be achieved by impersonal corporations and institutions - rarely individuals. In an effort to put this right as far as XML is concerned, Tim Bray, Sun Microsystems' director of web …
I Doubt We'll be Seeing the Back of XML for Some Time
Here's a nice testimonial:
"I think open source would have happened without XML and vice versa so they are not necessarily linked."
Is this guy for real? As has been mentioned countless times in the other thread relating to open source, OS did not start with Parens or any of the other jumped up little egomaniacs who are trying to snatch all claim for every historic thing on the net.
Back in 98, I distictly remember being at university where a mix of linux, unix and windows was in use. I was also dabbling with linux (yes, open source) at home at the time. The OSI is not open source. They are the self-imposed guardians of just one definition of what open source might be (to which the attach the label "The Open Source Definition" for added self-importance), and issuers of membership to their little clique. They did not start OS, and they certainly do not control it.
Now we are supposed to believe that nobody questioned the need for XML? Nobody ever wondered why a bloated CSV replacement would be needed when CSV runs just fine and with less parsing overhead thankyou very much?
It's human readable the evangelists claim. So's a sodding book, but shipping books between places is not a very efficient means of data transfer. It can support various formats using DTD, well CSV supported various formats by simply including the relevant data. About the only thing XML is good for is where branch sizes are unknown, and even then you'd get a lot less overhead with JSON, bencoding or some other format.
There are very few circumstances in which only XML will do, and even less in which XML is the most effecient method to use. RSS would have worked just as well, and with a lot less overhead in bandwidth, processing and programmer time using CSV. It may even have worked better given that a mispaced comma will cause the line to be skipped with any decent csv parser, a misplaced < can break XML parsing totally. Hell, even an unescaped & can cause clients to freak out.
You really don't understand what XML is, do you?
You think it's a replacement for CSV, a container for passing data between processes? Dear me, I hope you don't work in IT.
I work with XML every day; huge streaming quantities of it. I publish books and web sites and online courses and assessments from XML master files. And I only use one application of XML for the purpose.
I can understand why you might think XML was crap if you thought all it was was a CSV replacement for data passing; but if you knew anything about XML at all you would know that it isn't.
Like I say, I hope you don't work in IT.
I have to agree
XML is a rather over-hyped text markup format often used in situations where it is entirely unsuited to the data it is meant to represent.
"You think it's a replacement for CSV, a container for passing data between processes?"
ummmm... what else is data for? Except to pass between different programs.
Or do you like opening up those raw XML files in your text editor and bathing in its reflected human-(barely)-readable glory just for the heck of it?
I think the point is - XML has been over-hyped and over-abused as XML fans try to shoe-horn it into every single aspect of a system, even when it is *not* appropriate to use.
CSV is a very ill-defined format. Meanwhile those of us in the real world who need to work with data in a compatible way will get on with XML until the next acronym becomes popular.
I should hope the W3C know what XML is.
"Extensible Markup Language (XML) is a simple, very flexible text format derived from SGML (ISO 8879). Originally designed to meet the challenges of large-scale electronic publishing, XML is also playing an increasingly important role in the exchange of a wide variety of data on the Web and elsewhere."
Simple? Not according to Bray who said it is too big and complicated.
Designed for large scale electronic publishing (which implies more than just web)?
Not according to the 1.0 spec (http://www.w3.org/TR/1998/REC-xml-19980210)
"Its goal is to enable generic SGML to be served, received, and processed on the Web in the way that is now possible with HTML"
Important in exchange of data? Oh, isn't that what I was talking about? At the time that XML came about, CSV was still widely used. It's still got its uses today. XML could not be displayed in any reasonable way by standard web browsers at the time (XSLT didn't even exist then), so what was its purpose? It got touted a lot as a data transfer format. For a lot of data transfer tasks it is overkill.
Apart from the fact that the "Nobody ever wondered" comment was meant to be possible thoughts of those not accepting XML at the time, I also mentioned alternatives to XML such as bencoding and JSON. The main argument being that XML is bloated and can usually be replaced with a more effecient format. If you think I'm wrong, show me how what you use XML for couldn't be done in a more processor effecient way in JSON, bencoding or (god forbid) serialized arrays. The argument for human readability really doesn't hold if you are dealing with "huge streaming quantities of it".
If I genuinely thought XML was nothing but a replacement for CSV, would I have mentioned the handling of unknown branch sizes? It's pretty hard to have arrays of arrays with an unknown count stored in CSV, but this is why we have JSON and others.
I also suspect that in your list of things you do daily with XML, some are reliant on XSLT which is not XML. You use XSLT to "transform" your data, I use pretty much any coding language to "transform" my data. Your base XML though still only contains data. That makes it a data container, no matter the tools you use to display the data.
As for the comments about not working in I.T, here's the thing. If you attempt to label everyone with a different opinion to you as unsuitable to work in I.T, then you will find yourself with nothing innovative at all. You would not be sitting behind a nice modern computer being able to post on the Reg, because nothing would have changed since the first adding machines. If nobody was allowed a differing opinion, why would anyone have bothered to invent your precious XML in the first place?
I suppose I could make comments about sheep being crap as I.T workers, or using the best tools for the job, but I don't feel the need to label you incompetent just because you have a different opinion. Who knows, you may be the guy who convinces me that XML aint so bad.
NO ONE in the real world uses XML for "HUGE STREAMING" datasets. Its meant for small datasets.
Huge datasets are manipulated by binary database formats with SQL and will be for at least another thirty years as they have been for the past thirty-odd,
Huge sets of data in XML (or any text markup format really) is inefficient and SLOW.
I've been doing software engineering for twenty-six years and like to think I know what I'm talking about.
If anything, XML is a replacement for ASN.1. Which is an absolute dog to work with. As a generalised representation for structured data, it does a pretty good job I think.
XML Works For Me.
So Does JSON.
So Does Prioprietary Object-Oriented Protocol (POOP)
Whatever tool I need for the job. That's one of the kewl things about modern times. We have a lot of choices.
Ones that work get used and maintained.
Ones that don't get swept into the dustbin.
I've learned that screwdrivers make lousy spanners, but there's nothing wrong with a screwdriver because I can't undo my faucet with it.
Has it's uses. IMO I think there has to be something else could take it's place that would be much more human readable, and not such a pig to write code for, and where when you referred to an element or a node.
everyone would agree what one you were talking about...It gets used in many places without much thought as to whether it's actually the best thing to use.
What is "Shoulda" and why did it happen sooner?
unbelievable, some of you right on the button. some of you waaaayyyy offf.. fasts are people, XML is the number one way to store and retrieve data, i can prove it. c r a i g (a t) c l i c k n e t w o r k s . c o . u k. now this aint a sales pitch i wil show you an xml based engine kick the living crap out of an rdbms you can think of. try me.