Brassica SCA [v1.0.0]

[Darren]

Aye I just noticed there's a find feature in the rules section now, that's a nice touch!

[bradrn]

Aye I just noticed there's a find feature in the rules section now, that's a nice touch!

Thanks! I’ve wanted it for a long time, but never quite got around to building it until now.

[Lērisama]

Is there any way I've missed to force a category to have a certain value for a feature in the replacement, without copying the value from somewhere else?

Using these categories as an example

Code: Select all

categories
-voice = p t k s ? ? ? ?
+voice = b d g z r l m n
C = &&voice h

-long = a e i o u
+long = ā ē ī ō ū
V = &&long
end

If I want to delete coda n h with compensatory lengthing and with n voicing the following consonant, I think I have to write

Code: Select all

V n$voice#v C / V$long#v C$voice#v ;h isn't affected by this voicing
[{[V -long]} {[V +long]}] h / [{[V +long]} {[V +long]}] / _ C

but the latter is hard to read and easy to make a typo in
I could put h in -voice, ? in +voice to make it match, and write

Code: Select all

V h$voice#v / V$-long#v / _ C

But that messes up the previous change, and I'd rather not redo my categories for making one sound change a bit nicer. I'd like to be able to do something along the lines of

Code: Select all

V h / V$+long / _ C

But I don't think that is possible with current Brassica.

All this is based on the assumption that neither of these work

Code: Select all

V h / [V +long] / _ C ; replacement category doesn't match target, replaces already long vowels with �
V h / [V V +long] / _ C ; from the output, I assume the same as the above, with the +long deleting duplicate lexemes. I thought this used to work; has there been a change in how intersection works?

If there is a way to get something along these lines to work, I'm not aware of it, so please tell me

P.s. please don't take this as a criticism of Brassica; it is my favourite sound change applier by a country mile, and it is almost always incredibly intuitive, but also very powerful. It's just that I only have to ask about the few things that I don't fully understand yet, or have a very mild complaint about (I did just do the extra typing for the first example for deleting h, after all). Similarly please don't think I'm pressuring you into implementing something immediately, or even at all; I know you've just done a lot of work on a big update, and don't want to do anymore with Brassica for a while, but if you ever decide you do, please could you consider a nicer syntax for doing this

Edit: fixing typos

[bradrn]

If I want to delete coda n h with compensatory lengthing and with n voicing the following consonant, I think I have to write

Code: Select all

V n$voice#v C / V$long#v C$voice#vc ;h isn't affected by this voicing
;[{[V -long]} {[V +long]}] h / [{[V +long]} {[V +long]}] / _ C

but the latter is hard to read and easy to make a typo in

(Ooh, that first line is clever. Took me a moment to figure out what it does though…)

Honestly, I think I’d just write the second line as two rules:

Code: Select all

[V -long] h / [V +long] / _ C
[V +long] h / [V +long] / _ C

It’s more verbose, but easier to write and understand.

I'd like to be able to do something along the lines of

Code: Select all

V h / V$+long / _ C

But I don't think that is possible with current Brassica.

All this is based on the assumption that neither of these work

Code: Select all

V h / [V +long] / _ C ; replacement category doesn't match target, replaces already long vowels with �
V h / [V V +long] / _ C ; from the output, I assume the same as the above, with the +long deleting duplicate lexemes. I thought this used to work; has there been a change in how intersection works?

Correct, this doesn’t work (and I’m not aware of it ever having worked). The lack of a dedicated syntax for it is a particular sore point in the current feature system — I like your suggestion for the new syntax!

(Now I’m trying to remember why I didn’t implement this before finalising v1.0.0. I think I actually did make an attempt at it, but ran into difficulties with fitting it into the existing design. It’s low-priority anyway, since there’s ways to write it using the existing syntax.)

it is my favourite sound change applier by a country mile

Many thanks for saying this! I’m very happy to see that you like it enough to have spent this effort understanding its intricacies. Something which you’ve done impressively well, in my estimation.

(As for suggestions for new features, they’re always welcome!)

[Lērisama]

(Ooh, that first line is clever. Took me a moment to figure out what it does though…)

Thanks. I probably should have commented better. (And of course it's only now I notice the typo…)

Honestly, I think I’d just write the second line as two rules:

Code: Select all

[V -long] h / [V +long] / _ C
[V +long] h / [V +long] / _ C

It’s more verbose, but easier to write and understand.

Yeah, that's about what I was expecting. Thanks for confirming it though

I'd like to be able to do something along the lines of

Code: Select all

V h / V$+long / _ C

But I don't think that is possible with current Brassica.

All this is based on the assumption that neither of these work

Code: Select all

V h / [V +long] / _ C ; replacement category doesn't match target, replaces already long vowels with �
V h / [V V +long] / _ C ; from the output, I assume the same as the above, with the +long deleting duplicate lexemes. I thought this used to work; has there been a change in how intersection works?

Correct, this doesn’t work (and I’m not aware of it ever having worked). The lack of a dedicated syntax for it is a particular sore point in the current feature system — I like your suggestion for the new syntax!

(Now I’m trying to remember why I didn’t implement this before finalising v1.0.0. I think I actually did make an attempt at it, but ran into difficulties with fitting it into the existing design. It’s low-priority anyway, since there’s ways to write it using the existing syntax.)

Interesting about it never having worked, I must be misremembering. And yes, it's clearly not a priority, just a possible simplification (on the user end)

I’m very happy to see that you like it enough to have spent this effort understanding its intricacies. Something which you’ve done impressively well, in my estimation.

Thank you, but I think I should add the reason I can spend effort on the intricacies is that most things are really intuitive. That and long, boring bus rides

(As for suggestions for new features, they’re always welcome!)

Good to know, thanks

[Lērisama]

For completeness, I've just realised it's possible to do it in one line like this

Code: Select all

[&-long &+long] h / [&+long &+long] / _ C

(I keep forgetting featural categories are still categories). Not that it's an improvement on the first one in readability, but it is shorter.

[bradrn]

For completeness, I've just realised it's possible to do it in one line like this

Code: Select all

[&-long &+long] h / [&+long &+long] / _ C

(I keep forgetting featural categories are still categories). Not that it's an improvement on the first one in readability, but it is shorter.

Oh, true, you can do that. Not sure why I never noticed. (For normal categories it’s a bit risky because they can be different lengths, but of course this isn’t an issue for featural categories.)

(By the way, you don’t actually need the first ampersand, although I personally like to include it.)

[quinterbeck]

Congrats, Brad! Great work

Could explain how // ^+Str+Pri _ is operating in this example from the docs? (It's under featural categories there)

Code: Select all

categories
C = m n p t k b d g f s v z r l y w

+Str+None = a e i o u ə ɨ
+Str+Sec  = à è ì ò ù ə̀ ɨ̀
+Str+Pri  = á é í ó ú ? ?

V = +Str+None &+Str+Sec &+Str+Pri
; or V = &&Str - see below
end

-rtl -1 [+Str+None -ə -ɨ] / +Str+Pri
-rtl C* +Str+None C* +Str+None C* / C* +Str+Sec C* +Str+None C* / _ // ^+Str+Pri _

; dukɨtipəta → dùkɨtìpətá
; dukɨtipetə → dukɨ̀tipétə
; dukɨtipətə → dùkɨtípətə

I get that it's preventing the trochaic rule from applying after the primary stress but not how... The wildcard is operating from right to left because of -rtl, right? I couldn't find the // syntax described in the docs, so that's probably where I'm tripping up.

I have a vague sketch with mad stress-based syncope that I want to try out in Brassica, so it would be useful to get my head around this

[bradrn]

I couldn't find the // syntax described in the docs, so that's probably where I'm tripping up.

Oh wow, that’s a pretty big omission. I can’t quite believe I forgot that.

(Luckily, the complete Reference Manual does mention it: https://github.com/bradrn/brassica/blob ... nge-syntax.)

In short, // is syntax for an exception. The sound change will not be applied in a position where the exception matches.

[quinterbeck]

In short, // is syntax for an exception. The sound change will not be applied in a position where the exception matches.

Aha! Great, thanks. I hadn't thought to look in the reference document yet

[Raholeun]

I am having trouble accessing the online Brassica. Is that just me?

[Raphael]

I am having trouble accessing the online Brassica. Is that just me?

Apparently it won't load for me, either.

[bradrn]

I am having trouble accessing the online Brassica. Is that just me?

We switched ISPs to my house, so perhaps that might have something to do with it. I’ll keep on investigating and let you know.

In the meantime, well, this is why there’s a desktop version… (if you’re not on OSX, at least. If you are, I can’t help you, because I have no way of building stuff for OSX.)

[Lērisama]

I am having trouble accessing the online Brassica. Is that just me?

We switched ISPs to my house, so perhaps that might have something to do with it. I’ll keep on investigating and let you know.

In the meantime, well, this is why there’s a desktop version… (if you’re not on OSX, at least. If you are, I can’t help you, because I have no way of building stuff for OSX.)

Good to know. I also can't load it, for what it's worth. (I also couldn't load it for around a day last week, but it went away, so I didn't mention it)

[bradrn]

I am having trouble accessing the online Brassica. Is that just me?

We switched ISPs to my house, so perhaps that might have something to do with it. I’ll keep on investigating and let you know.

In the meantime, well, this is why there’s a desktop version… (if you’re not on OSX, at least. If you are, I can’t help you, because I have no way of building stuff for OSX.)

Good to know. I also can't load it, for what it's worth. (I also couldn't load it for around a day last week, but it went away, so I didn't mention it)

Yeah, it’s still not working, sorry. It looks like it is indeed related to the ISP: apparently it blocks incoming HTTP and HTTPS connections. I tried one or two small things, but it looks like they aren’t working; I may need to find another way entirely to host it.

[Travis B.]

We switched ISPs to my house, so perhaps that might have something to do with it. I’ll keep on investigating and let you know.

In the meantime, well, this is why there’s a desktop version… (if you’re not on OSX, at least. If you are, I can’t help you, because I have no way of building stuff for OSX.)

Good to know. I also can't load it, for what it's worth. (I also couldn't load it for around a day last week, but it went away, so I didn't mention it)

Yeah, it’s still not working, sorry. It looks like it is indeed related to the ISP: apparently it blocks incoming HTTP and HTTPS connections. I tried one or two small things, but it looks like they aren’t working; I may need to find another way entirely to host it.

Maybe host it on github.io? That's what I do with zeptocom.js. Also you could make it trivial to host locally with git clone followed by python3 -m http.server if you haven't already.

[bradrn]

Well, I seems to have got it working, though via HTTP only: http://bradrn.com/brassica/. I’ll have to figure out how to get HTTPS working again.

[Darren]

Well, I seems to have got it working, though via HTTP only: http://bradrn.com/brassica/. I’ll have to figure out how to get HTTPS working again.

Still not working for me I'm afraid.

[bradrn]

Well, I seems to have got it working, though via HTTP only: http://bradrn.com/brassica/. I’ll have to figure out how to get HTTPS working again.

Still not working for me I'm afraid.

Strange… it works for me. In more detail, what do you experience? (Error message or other symptoms… but then, you know this stuff as well as I do, I’m sure you can tell me what I need to know.)

[Raholeun]

Sadly, the linked page does not load for me either. It states there is a connection time-out: "The server at bradrn.com is taking too long to respond."