My reason is written in my latest comment here:
https://github.com/lojban/jbovlaste/issues/184

I just repeat what I wrote below the line:

--------

lagleki, you do not understand. People are free to choose any word form
they like.
Because, according to CLL, all forms a lujvo are equivalent by definition.
So if I define ?tolyli'a?, I also automatically have defined ?to'e zei
cliva? and ?tolycliva? as well.

Therefore, it does not make much sense to include all those word forms,
don't you think? The other word forms are simply implied.

That the dictionary always should use the canonical form is simply to avoid
redundancy.
It would be bad if we had ?tolyli'a?, ?to'e zei cliva? AND ?tolycliva?
floating around. It's simply a technical reason.

That also why there is ?malgli? in the dictionary, although ?malglico? is a
much more common (and 100% equally valid) form of this lujvo.

In fact, this is stated every time you create a word:

If it is a lujvo, make sure you're using the lowest scoring form of the
lujvo. You're not going to get anywhere by using a different form, as the
database is periodically swept for noncanonical form lujvo, and they are
modified to the appropriate canonical form. (Or removed, and all related
data pointed towards the canonical version, if it already exists.)

I wonder if this database cleanup has ever been actually done before.

It would be even better if Jbovlaste could check the lujvo score right from
the beginning.

--------


PS: People not being able to find alternate lujvo forms is actually a
software problem that needs to be fixed in the software. The software
really should be able to show all equivalent lujvo forms at once. That
could all be done automatically. Having multiple entries for basically the
same lujvo is bad. :-(