~bleep.txt (NSFW!)

[DeusExMachina]

A quick, idea, one I think is easy-peasy.
Currently, bleep text works by subbing in asterisks for the middle chars. Mostly this works fine. But I think even more utility can be wrung from this if instead of just a flat list and a replacement rule, it be made even simpler, with a just a substitution list.
So instead of a flat list, it would contains pairs, the first item in the pair being the target word, and the second item being the replacement. So the current file which might read:
fuck
shit
cunt
bitch
etc.
would read:
fuck,f**k
shit,s**t
cunt,c**t
bitch,b***h
etc..

The advantage here is that you can then do far more powerful subs, like for instance:
cuz,'cause
yeh,yeah
yea,yeah
cos,'cause
gunna,gonna
etc.,
and all those other annoying misspellings
and even
...,…
and such

[dave]

great idea!

[DeusExMachina]

Just a thought:
With a well placed if-statement, we could leave the current functionality in place, and add the above-mentioned capability.
Then the format could be:
Single item on line, use asterisk replacement routine
More than one item, replace any instance of the initial members of a line with the final member of the line. E.g., the following ~bleeb.txt file:

bitch
cuz, 'cuz, cos, 'cos, 'cause
shit
yeh, yah, yeah

would result in the following substitutions:
b***h
'cause
s**t
yeah

I suspect this should be just a minimal edit.

[dave]

actually that would be more work.
much easier if it's one substitution per line, like this:

Code: Select all

bitch
cuz, 'cause
'cuz, 'cause
cos, 'cause
'cos, 'cause
shit
yeh, yeah
yah, yeah

[DeusExMachina]

Either way works fine. Just thinking of data file efficiency and redundancy minimization (old computer science class habits!).

Question: how hard would it be to mark any blobs where such a substitution was made, say render in red instead of the default purple.
Just an idea. Mostly it occurred to me because of the issue with words like "ass" being subbed in the middle of words, leading to things like "a*sumption" and "pa*sing". Having them marked in the timeline would make checking for such false subs much simpler. Know nothing about the architecture here, so no clue if this is relatively simple, or a royal PITA that would have to wait for v2.0/"The GarageBand Interface edition".

Oh, BTW, as mentioned on the Bugs page, the conjoined words problem is definitely a problem that only effects the first two words of a line. Still see no common trait that links all the instances other than that, so pretty sure there is some underlying math error or a*sumption that is leading to placing the words directly abutting each other.

[dave]

pleeeeeze: separate subjects should get separate posts.

not hard, just time consuming.