At 12:33 AM 3/9/99 -0600, Jason L Tibbitts III wrote:
>I initially rejected this approach because I didn't want to worry about new
>commands breaking common abbreviations by making them non-unique. I think
>it's probably worth revisiting, especially since someone gave me code to do
>it. (Sometimes code is better than money for the purposes of bribery.)
It would be worth noting in the documentation that use of abbreviations
has the risk of becoming non-unique at some point in the future, so
abbreviations are best for interactive use and best AVOIDED in scripts.
>Yes, that is reasonable. Fortunately we have the hooks in place to do this
>already. I think some restriction on the minimum length of an abbreviation
>is a good idea. (A one or two character command is probably a typo.)
The code I submitted has commented-out error text (since I didn't
know the right way to present it in the error message Mj2 sends back
to the user) - see my post on March 6th. If the abbreviation is not
unique, it lists all available commands and aliases (40 or 50 of them)
which is something that might stop some user questions before they are
even asked. Right now the server says it's illegal, but doesn't suggest
what to write instead. In the case of abbreviations, it might be
appropriate to list only the commands which match, instead of all
commands.
I agree that a minimum abbreviation of 3 characters is wise... but
that was not in the code I submitted because I didn't think about it.
References:
|
|