# Cross Browser Compatible XSLT

I recently found myself inclined to learn
[XSLT](https://www.wikipedia.org/wiki/XSLT). I find the best way to learn a
language is to apply it to something practical, so I used it to update my <abbr
title="Curriculum Vitae">cv</abbr>/resume.

The interesting thing about XSLT is that some browsers support client side
transformations, whilst other browsers don't. I wanted to take advantage of the
capability of the clients which support XSLT whilst making it compatible with
the others.

Using [Apache HTTPd](https://httpd.apache.org/) as the web server, the solution
I came up with was to serve XML to those clients which support XSLT, and to do
the transformation with [mod_xslt](http://www.mod-xslt2.com/) otherwise.

I had to do a fair bit of research to get a list of user agents of browsers that
support client side XSLT. I'm sure the list I came up with isn't comprehensive,
but it should cover 99% of the web browsers in use. The worse that can happen is
a transformation is unnecessarily done server side when it could have been done
client side. I came up with the following Apache configuration which I dropped
in my .htaccess file:

```apache
## Serve index.xml by default
  DirectoryIndex index.xml

## Set server side transformation for xml responses
  AddOutputFilterByType mod-xslt application/xml

## Set client side transformation for:
  BrowserMatch \bMSIE\s no-xslt
  BrowserMatch \bFirefox/\d no-xslt
  BrowserMatch \bSeaMonkey\d no-xslt
  BrowserMatch \bSafari/[3-9]\d\d no-xslt
  BrowserMatch \bOpera[/\s]\d no-xslt

## Turn it back to server side transformation for:
  BrowserMatch \bMSIE\s[1-5]\. !no-xslt
  BrowserMatch \bOpera[\s/][1-8]\.\d !no-xslt
  BrowserMatchNoCase (bot|spider) !no-xslt

## Make sure proxies don’t deliver the wrong content
  Header append Vary User-Agent
```

You might wonder why you would use XSLT for something like a CV. Well, all I
have to do now is write an MS Word stylesheet and a PDF stylesheet and I'll be
able to update my CV in all three formats xhtml/doc/pdf by updating a single XML
file. Neat huh?

**UPDATE:**

I've given up on client side XSLT for now. There are too many quirks and
problems with it and it is just safer to do the translation on the server side.
A few problems:

1. Firefox doesn't support disable-output-escaping

2. The NoScript Firefox addon disables XSLT as well as JavaScript

3. Old versions of Netscape don't support importing external xslt