I am in full agreement that using file extensions to indicate content
type is quite reasonable, but there are still a few reasons why I am
hesitant to use them in Roller ...
1. .html and .xml typically indicate static content to me and we are
talking about using them for dynamic pages. technically there is
nothing wrong with that, but i think some people consider that confusing.
/<weblog>.html?cat=<category>
/<weblog>/feed/atom.xml?excerpts=true
2. adding .<file-ext> into the url *could* pose an additional parsing
problem. right now Roller doesn't allow "." in a weblog handle and
various other user defined components of a url, but if that restriction
was ever lifted or not properly enforced then we would have parsing
problems. someone with a blog named "foo.html" or and entry anchored
"blah.html" might trigger a parsing problem. that may be too unlikely
to really be considered a concern, but it came to mind.
3. file extensions just look a bit ugly and unnecessary to me. if i am
given this set of urls to choose from i would always pick "/login" as
the best ...
/login.jsp
/login.html
/login
-- Allen
John Hoffmann wrote:
Dave,
I like them if they indicate content type and not the implementation.
Implementation to me means advertizing the tool used to make a content
type and is totally irrelevant to a URL: php, jsp, asp etc. html is a
content type.
Here is an admitedly older article about page extensions and google rank:
------
**http://www.seotoday.com/browse.php/category/articles/id/125/index.php
*Dynamic Pages*
This is another issue that is a gamble. Google will index file types
including html, pdf, asp, jsp, hdml, shtml, xml, cfml, doc, xls, ppt,
rtf, wks, lwp, wri. The problem with this is certain file types and
extensions are more likely to rank better and are more optimizable.
Google has officially stated that it will only index a certain amount of
dynamically generated pages, which includes URLs with a query string, as
a result of server issues....
------
I know google does lots of computation on the URL - that is why putting
those tags in the URL will be great :) Wether or not they still base
content type of of extension or metadata I don't know - it used to be
that they threw out all metadata since it was heavily spoofed.
So overall - I would be reticent to throw away extensions - I think they
add clarity - and Google may too...
-John
David M Johnson wrote:
On May 3, 2006, at 4:49 PM, John Hoffmann wrote:
Yes, you hit on a pet peeve. Extensions for content type not
implementation.
When designing the JavaOne web site 5 years ago we made every page
available in 4 formats which was controlled by the extension.
standard ones:
javaone/2001/session-1234/detail.html (for desktop browser visitors)
javaone/2001/session-1234/detail.xml (for crawling by 3rd party
data harvesters)
two custom types:
javaone/2001/session-1234/detail.lite (for small screen devices)
javaone/2001/session-1234/detail.prt (for printing)
John, are you saying that you are for or against file extensions?
- Dave
