Hacker News Reader

Comment by bibonix | Hacker News Reader

XSLT is one of the most powerful and elegant technologies created in the last 15 years. Those who don't understand it and that's why can't use it, should just do their home work and learn better

View on HN

Since two years I am working full time on a website which gets rendered via XSLT (via Apache Cocoon). When I joined, it was kind of a mess and I had a hard time understanding XSLT and the templates written.

Just recently I had the chance to do a rewrite. What I did is I created "my own HTML". Basically every module of Twitter Bootstrap has it's own XSLT template. That means you have a very easy XML "HTML" syntax, but the output is Twitter Bootstrap. And every piece of output HTML is defined just once in the whole application so it's easy to maintain.

With the help of XSLT you can abstract a lot of things. One example: I have an element called <colgroup/>. It can contain up to 12 <col/> elements. if I set the @size attribute to one of the columns, the @size attribute for the others will be calculated automatically and the output matches the Twitter Bootstrap CSS classes.

I have to say, I love it. I can't imaging writing the whole mess of Twitter Bootstrap plain HTML in a template anymore.

My definition of powerful and elegant is lisp

XML is just death by overengineering

apples and oranges? There is no other way of encoding structured documents these days than XML. Like it or not, XML is the de facto standard for data exchange (export from databases, product information software,....)

XML might be overengeneered (which, except for a few things I don't agree with), but there is currently no alternative for it.

The site in question lists several alternatives, some of which are widely in use especially in cases where XML falls short: brevity, fast to parse, easily readable... Most of all, calling it the de facto standard is either dishonest or clueless.

I am working for many years in the publishing business (dealing with structured documents, product data etc.). I can tell you, that all of my customers are more open XML than any other document formats (there exist none as suitable for the job).

You can call me clueless or dishonest, I don't care. I can only share my experience with the topic. You don't have to believe me.

Your customers in the field you work in, perhaps, but it's hard to tell that's what you mean when you write data exchange. It's a wide field that is not limited to the publishing business or your customers.

> Most of all, calling it the de facto standard is either dishonest or clueless

It is in fact the de facto standard in the publishing industry. The "other" format is of course PDF.

Maybe all of this will change when we have more technologies that support formats that can handle mixed content as easily as XML.

Does anyone use Microsoft XPS? It looked rather interesting; like PDF without any interactivity (except for links) and with special support for publishing (color management, job tickets, etc.). And internally it is a collection of XML docs and binary data zipped together into a single file; pretty neat, must be easier to use in automated workflows.

I can't answer the question but anyway:

None of my customers have ever asked for that. I have not seen a printing house that demands XPS. So I doubt that it plays any role in the market (Germany here).

Thanks, that's informative. I haven't seen it used much either; the only real use I saw (aside from viewing) was an XPS printer driver for a Canon inkjet printer. (I've also heard that it produced better results than the standard driver, but have no first-hand experience.)

databases are not structured documents, you can export them eg in JSON perfectly well. Or for that matter in SQL as is the usual practise. Marked up structured text is a different matter, XML still has a use case there.

I was a bit unclear. When I get database dumps, I get them as Excel files or as an XML document.

Most of the times the documents I get are hierarchically structured.

Yes, JSON could be fine as well. But it simply lacks a standard toolchain which XSLT ans its ilk proides.

I am not trying to defend XML in any way. I just want to say the two things:

a) my customers never deal with JSON, but often with XML, so JSON (and other formats) are not an issue for me b) There is a very nice toolchain for XML, including formatters, tranformation tools, database publishing tools (my very own: https://speedata.github.io/publisher/index.html) and many others. I have not found such a toolchain for other formats.

But XML does not map well to a database dump. A database is a set of sets of tuples. It is not generically hierarchical. It maps to a set of csv files say, as a transfer medium, but it is designed to be manipulated through relational calculus, which is not easily mapped to xslt.

You are using tools not because they map well to the problem space, but because they are the tools you have, which your customers want, and which you are familiar with, but that does not mean they are actually mapping well to the domain.

Actually the XML part is the one that maps best to the problem space. I deal with structured documents (such as part/title/paragraph ... or product group/product/components). The database part (usually Excel sheets or SQL based stores) are the ones that are "insufficient".

> XML might be overengeneered (which, except for a few things I don't agree with), but there is currently no alternative for it.

There's perhaps no general alternative to it that covers all of the things XML tries to do; there are lots of specific alternatives that cover specific things that XML tries to do. The complaint against XML isn't that there is a better general replacement so much that there is a better replacement for each (or at least, very many) of the applications and that trying to shoehorn all of them into a single solution has costs that outweigh the benefits.

   > My definition of powerful and elegant is lisp

Dude, XML is just s-exprs and XSLT is macros.

REALLY? http://en.wikipedia.org/wiki/XSLT#Example_1_.28transforming_...

I want to enter this piece of evidence into the trial:

Comma separated string parsing XSLT: http://stackoverflow.com/a/2850181/92493

If you are using XSLT to work on anything other than XML you are doing very strange things.

Oh good, so it's only good in the world of XML

That's what I call USELESS

XML and XSLT are analagous to s-exprs and macros, to say they are "just" those things is willful ignorance of a whole boatload of complexity.

Good, so I'll replace the '<' in XML with '#:-D' and the '>' with '%:-/'

Or replace indenting levels with the XPath of where it belongs and make order optional

Syntax matters.

Tell any lisp programmer he doesn't get first-class functions and has to use nothing but macros, and he'll run screaming in horror.

> XML is just s-exprs

No. That myth has been decisively addressed by Erik Naggum about 12 years ago. His summary:

> They are not identical. The aspects you are willing to ignore are > more important than the aspects you are willing to accept. Robbery > is not just another way of making a living, rape is not just another > way of satisfying basic human needs, torture is not just another way > of interrogation. And XML is not just another way of writing S-exps. > There are some things in life that you do not do if you want to be a > moral being and feel proud of what you have accomplished.

Please read his posting/rant for the arguments. Dude. (I'll just tell you to search for "naggum xml", there are more than enough copies in circulation, and you'll find a few more postings by other people.)

Now, as for XSLT: The big problem is the hairy syntax. It is really (at least) two languages (the XML tags, and the query language that is used inside selectors). In effect, you are writing at least three languages completely intermixed in a single file: the output language (most often some XML or HTML variant), the XSLT tag language (another XML format), and the XSLT query language (an incredibly limited ad hoc micro-language inside some XML attributes).

XSLT is a very limited language, as opposed to Lisp macros, which can use the entire Lisp language.

And yes, I have used XSLT in my job, and I do have reason to think that the XSLT-stylesheets I wrote have an acceptable quality. However, I know that I could have done their job better if I could have used some structured data, an HTML formatter, and a real programming language.

The problem is that it's often misused. Using an XSLT that takes 0.5s to run to transform some XML to HTML during web requests on a busy site? Idiotic.

Using it for async transformations - html to pdf, customer message format to your message format. Fine.

The problem is not that it is misused. The problem is that there is a small subset of problems for which it works passably (like many other technologies) and a large universe of problems for which it will cause you massive pain.

Using compiled XSLT-translets in Java (XSLTC) has a huge impact on performance, with transformations taking 0.5s the problem seems to be in the way your XSLT is written and pushing or pulling the data, not XSLT itself: http://xalan.apache.org/old/xalan-j/xsltc_usage.html

Compiled XSLT stylesheets must be very efficient; they're automata, like regular expressions.

Reply on news.ycombinator.com