On Jun 7, 2004, at 10:47 AM, David W. Fenton wrote:

Seems to me that Lyrics got the way it is because it is an *old*
subsystem, dating back to very early versions of Finale, and the
changes to it have been bolted on the sides over time, making it
rather baroque and nearly impossible to figure out. And the problem
is that the basic system on which it is all built does not have the
best architecture to support all these additional features.

It is exactly the kind of component that appears to me to need a
major rewrite, from the ground up, so that the basic architecture
could be altered so that all the parts that have been bolted on could
be integrated into the basic structure, instead of having this kludgy
afterthought feel to them.

I think you're exactly right on that.

As far as I can tell, the underlying structure of Lyrics is modeled on that of Expressions. That is, your lyric data is really like a gigantic expressions list, so when you attach the syllable "of" to a note, you're adding the word "of" to the list and attaching a pointer to it to the note.

There's a certain logic to that approach with regard to expressions, where it really does make sense to think of every "ritard" as a single definition which might appear numerous times, but there's very little to gain from it with respect to lyrics. It's pretty rare that you'd want multiple assignments for the same lyric at all. Even if "of" appears several times in your lyric, you're still going to have it appear multiple times in the list, so what's the point of using indirection here?

In that respect, it seems like it would be more logical if there is no syllable list and each syllable is simple in and of itself an attribute of the note, as it is in other engraving programs. That sort of structure would be a natural fit with the Type-in-Score input method. The lack of such a structure in Finale is what makes Type-in-Score have such a kludgy, bolted-on feel to it (and probably is also why it arrived relatively late in Finale's history).

The key is to make sure that in switching, you don't give up the advantages that the old system offers: the features that derive from the ability to identify a collection of syllables as a group (verse), and the ability to identify one syllable as "following" another (ie, cut-and-paste access to the lyric text in its own window, ability to change a verse's font/size or baseline en masse, ability to shift lyric assignments). But those labelings could be attached to the syllables under the structure, and it would be smoother than the other way around.

Indeed, a lot of very small changes could make Lyrics much more
usable (like allowing resizing of the click assignment dialog --
geez, how frigging hard would *that* be?),

Amen! And the Edit Lyrics window, too. Why on earth are these windows stuck in the mid-1990s, UI-wise? That's just embarrassing.


 but the basic underlying
problem of the Lyrics tool, that the visual representation of the
data and the actual underlying data do not have a clear relationship
to each other (and, secondarily, that the UIs involved require you to
understand the relationship between the displayed data and the stored
data), could, I think, only be rectified by a ground-up rewrite of
the whole thing.

mdl

_______________________________________________
Finale mailing list
[EMAIL PROTECTED]
http://lists.shsu.edu/mailman/listinfo/finale

Reply via email to