Friday, March 7, 2008

Of domains and filenames…

Quick and dirty overview:

When you make a link to a page on our library web site, use the full domain name:


When linking to an index (or default) file within a folder (often index.html) and you omit the index/deafult file name to shorten the URL, make sure to terminate the URL with a trailing slash.


Okay, a bit more for those interested (or looking to while away a Friday afternoon):

I just wanted to make a quick post about URLs for library web pages (I am deep into finishing up our ARL Spec Survey right now, so I'll make a longer regular update this weekend). There are a myriad of external policy influences that play into this, as well as internal systems issues at play (which I will detail in an upcoming best practices documentation in the CMS and off my WTCC web site) but what it boils down to is this: when you are preparing communications that include links to the library, keep two things in mind. First, the proper domain name to use is (currently, you never know when we may have an external policy decision that changes this):

Although library ISD has done their best to catch all variations of this, like the most common one, (and trust me, there are many more variations, and I'm not going to list them, mostly because I don't want the mere mention of them to accidentally perpetuate their use) we all should really use from now on. Although in some (maybe even most) cases a variation of this domain will work, to ensure that it will 100% of the time, regardless of if your content is on the the old web server or the new CMS, use the full domain, .

A related issue arises when we publicize truncated URLs when people are linking to the index.html page within a directory. For instance, let say we have a top level folder called "myfiles" and an index.html page within that folder. The proper URL to publicize in that case is:

Again, ISD has implemented several helper applications on our various web servers to push variations of this (like ) to the correct file, but not all of the systems we use will catch this 100% of the time (again, look for future best practices documentation for the long-winded and technical explanation of why this is), so please do not use these incorrect approaches in the future just because they _seem_ to have worked in the past.

On a related note, it's a good practice to have your web sites/folders/pages posted for testing before distributing documentation that includes URLs, even is this means simply putting up a page stub that you'll complete later.