4 ms·

indeed, neat e.g. for nicely (with modifiable styling) viewing immutable raw data like messages: https://mro.name/microblog https://mro.name/microblog Have mu

indeed, neat e.g. for nicely (with modifiable styling) viewing immutable raw data like messages: https://mro.name/microblog https://mro.name/microblog Have mutable design on immutable data is awesome for long-term concerns. And having the data raw!

giantrobot 4y ago
The raw data + styling model is ideal to me. You've got one data endpoint for both people and machines that's equally useful to both.
- mro_name 4y ago
  I wonder why it's seen so rare. Are people afraid to disclose the data? Ignorance about the possibility? SEO?
  - giantrobot 4y ago
    It's a little bit of all of those factors. In the early days of XSLT client side support was limited. IE still dominated the web in the early 2000s. IE 5/5.5 only supported the XSLT working draft and made up more than half of IE's market share. In 2001 if you were excited about client side XSLT it was only available to fewer than half of your visitors at best. You could do XSLT on the server side and spit out HTML but that was just one of several ways of converting your raw data to HTML. For most developers it was easier to go from a SQL database through ASP, PHP, or whatever to HTML. So in the early days it didn't gain much momentum. By the time browsers supporting client side XSLT were dominant AJAX was pretty well established as the new hot thing. JavaScript was ingesting XML with little interest in having the browser handle it directly. The tooling in the early days wasn't that good either. Dreamweaver 8 added support around 2005 but I think by then the AJAX train had sailed. Before then developing an XSL stylesheet meant a lot of rerunning a processor and reloading a page. For the same effort you could just do the work in PHP or ASP and get more programmability. Today everyone has a browser capable of doing client side XSLT very efficiently but XML is seen as old and crusty. No one wants to go back and learn XSLT because most developers seem to have doubled down on JavaScript. Since XML as an endpoint format has been ignored for so long I would bet it doesn't have a good SEO story. While it can carry all the metadata Google or Facebook wants, if they don't crawl it then it doesn't matter.

giantrobot 4y ago

The raw data + styling model is ideal to me. You've got one data endpoint for both people and machines that's equally useful to both.

mro_name 4y ago
I wonder why it's seen so rare. Are people afraid to disclose the data? Ignorance about the possibility? SEO?
- giantrobot 4y ago
  It's a little bit of all of those factors. In the early days of XSLT client side support was limited. IE still dominated the web in the early 2000s. IE 5/5.5 only supported the XSLT working draft and made up more than half of IE's market share. In 2001 if you were excited about client side XSLT it was only available to fewer than half of your visitors at best. You could do XSLT on the server side and spit out HTML but that was just one of several ways of converting your raw data to HTML. For most developers it was easier to go from a SQL database through ASP, PHP, or whatever to HTML. So in the early days it didn't gain much momentum. By the time browsers supporting client side XSLT were dominant AJAX was pretty well established as the new hot thing. JavaScript was ingesting XML with little interest in having the browser handle it directly. The tooling in the early days wasn't that good either. Dreamweaver 8 added support around 2005 but I think by then the AJAX train had sailed. Before then developing an XSL stylesheet meant a lot of rerunning a processor and reloading a page. For the same effort you could just do the work in PHP or ASP and get more programmability. Today everyone has a browser capable of doing client side XSLT very efficiently but XML is seen as old and crusty. No one wants to go back and learn XSLT because most developers seem to have doubled down on JavaScript. Since XML as an endpoint format has been ignored for so long I would bet it doesn't have a good SEO story. While it can carry all the metadata Google or Facebook wants, if they don't crawl it then it doesn't matter.