You lose...

At pretty much the same time as my last posting, I was involved in a discussion of how our internal Document Management System was going. As happens with these things, discussion had gotten rather heated over the nitty gritty. In particular, the document file naming conventions had a few glitches to iron out, and was raising my ire.

The standard being using calls for the project names and official release numbers to be included as prefixes to the file name. Apart from the bizarre but acceptable notion of continually updating the filename (a hangover from not so olden times, when files were being stored in a flat directory) I took exception to the use of prefixes. My argument went like this:

• The really useful, standout bit of a filename is its verbally descriptive nature.
• Adding a monotonous prefix (with unique index number) hides this information from casual browsing
• It is difficult to sort on name usefully for the same reason
  
• Put the indexing as a suffix, if you must have it in the name

Consider how it would look if I started all my blogs with the formula 'Randomised_nnnn_'. Very spontaneous!

So here we have (I think) a reasoned argument. It lists objections, why I object, and alternatives.
The response ( a depressingly familiar one, alas!) was along the lines of:

• we do it this way
• the standard requires it
• the standard does not get changed. (axiomatic)
  

Oh well! Can't win them all! Maybe I should learn to recognise windmills from a bit further away!

However, since this is my blog, I get to have the last word. So, consider this:

• A standard's business is to state what is to be done
• A good standard will also discuss why it is to be done
• A bad standard will stipulate how it is to be done