Why the rush?
Well, I think it's quite major. We should work this out more before making any changes, because it has a wide effect on many templates, not just the linking templates. It's a complete change in standard practice across Wiktionary. For example, it would also affect templates like {{form of}}, {{suffixcat}} and others. Anything that's used to link to another entry in one way or another.
Personally, I'm not decided whether a change is even desirable. The nice thing about {{l}} and the likes is that the two term parameters mimic how plain wikilinks work. {{l|la|ordō|ordō}} somewhat resembles [[ordo|ordō]], which is a bonus for those who understand wikicode and only need to learn Wiktionary's practice. It makes the practice relatively simple: take the existing plain wikilink, remove the square brackets, and wrap it inside {{l|(lang)|......}}. However, if we are to change anything at all, I think the best whould be to move the gloss down one parameter, and use alt= for the display form.
In any case, renaming {{term/t}} and {{term}} to {{m}} doesn't really have anything to do with that, because it only affects one template and does not affect any proposed change in the parameters.
It kind of does, becauſe when the new ſyntax is agreed upon, we will have to rewrite all direct tranſcluſions of "mention" templates all over again. So we might as well do it in one go.
And wiþ some care, the whole change does not have to be ſuch a flag day eiþer. I added a ſeparate "incompat" mode ſpecifically to give it chance to be teſted in a ſmall ſcale before changing all templates.
And by þe way: I would like to ſee a |nongloss= parameter (or ſimilar) added to templates for annotating terms with non-gloſs definitions. Say, {{m|pl|-ek|nongloss=diminutive ſuffix}}.
You're thinking of it in time and effort terms. But it's almost no effort to run a bot twice, and time isn't really an issue either is it? In fact, to change the parameters around we would need to create tracking categories/templates first anyway, to feed such a bot.
And {{l}} has a pos= parameter which does what you describe, more or less.
Not really. I want it tested for a while so that we can see it in some real situations and determine whether this is a good idea at all. Or simply to make more people aware of the idea before suddenly converting everything at once and making it a fait accompli much to their dismay. Like we did with {{label}}, with, as I think, some success.