XML is based on HTML. (Or, more pedantically, HTML and XML are both 
based on SGML.)

HTML is a *markup language*. It takes a big block of text, and inserts 
one or two small marks to indicate section headings, hyperlinks, text 
emphasis, and so on. But basically it's a textual document, just with a 
few formatting marks.

XML is intended to be a universal way to represent all data. Which isn't 
nearly the same thing. It's called eXtensible Markup Language, but it's 
really eXtensible Container Format.

It is no secret that human-readable file formats tend to be much, much 
more portable (and extensible). Suddenly you don't have to deal with 
things like big-endian vs little-endian, signed vs unsigned, 32-bit vs 
64-bit, etc. (No, instead you have to deal with the details of ASCII 
number formatting, e.g., is ".5" acceptable? Or must it be "0.5"?)

It's also no secret that textual formats are less efficient. (But hey, 
it never stopped PostScript!)

The nice thing about XML is that, since it's a standard, anybody that 
wants to can make up some format based on XML, and then anyone who 
understands XML has some small chance of figuring out what it all means. 
There are standard XML parsing and processing libraries. There are 
standard tools for searching, sorting and transforming XML into other XML.

The problem is... sometimes XML isn't a good fit. From what I can tell, 
SVG works reasonably well. But something like MathML is... impossible to 
read or write by hand. It's just absurd. The format is clearly and 
obviously designed for ease of machine manipulation, not for humans.

There is also the minor detail that XML is actually quite a lot more 
complex than most people realise. Most people think that "XML" just 
means "write stuff in little angle brackets". In fact there is much, 
much more to it than that. It's quite unecessarily complicated, in fact!

Post a reply to this message