| Search Results |
Total results: 10 |
| 1 |
Versatile Sound Change Applier
/ Re: Position switching using categories? |
on: May 15th, 2007, 06:36am |
| Started by Contrail | Post by MUBA |
Sorry for being a bit late noticing your question.
But the answer is really simple.
You see,
VLS = ptk
VS = bdg
S = <VLS><VS>
[SL]/[LS]/_
This creates three variabeles, namely <VLS>, <VS>, and <S>. But in the rule you also try to use the variable <L>, which has never been defined. So you probably want to put an
L=lr
line before the rule, too.
Hope this helps
|
  
|
| 2 |
Versatile Sound Change Applier
/ Position switching using categories? |
on: May 5th, 2007, 09:07am |
| Started by Contrail | Post by Contrail |
Hi, I've just recently started using VSCA and so far it's working well for me, but I've run into a small snag trying to do position switching (such as changing rk to kr) for categories of sounds.
What I tried doing that didn't work was the following:
# Voiceless Stops
VLS = ptk
# Voiced Stops
VS = bdg
# Stops
S = <VLS><VS>
# Swap LS for SL
[SL]/[LS]/_
What I ended up getting from that was things like lb changing to kl, not bl like I intended. Splitting each stop into it's own rule such as [bL]/[Lb]/_ works, but I'd like to know if there's a way to do it without so many seperate rules.
- Contrail
|
|
|
| 3 |
Versatile Sound Change Applier
/ Re: UTF8 error |
on: Apr 29th, 2007, 5:12pm |
| Started by Jashan | Post by Jashan |
Spoke too soon. I put the (C)(C)(C) back into the now "fixed" rule variant, and the problem returned. I tried reducing it down to just one (C) and still no luck.
(At this point I'm working with a partial rules file, adding rules 2-3 at a time to ensure they're working, so it's not having to process the entire script.)
|
|
|
| 4 |
Versatile Sound Change Applier
/ Re: UTF8 error |
on: Apr 29th, 2007, 5:10pm |
| Started by Jashan | Post by MUBA |
Thank you 
At the moment I don't have the energy nor the time to find out why it causes a memory-hang, but I intend to solve it another time.
Just a question for you: now that you changed the rules to avoid the problem, is it still accurate enough?
|
|
|
| 5 |
Versatile Sound Change Applier
/ Re: UTF8 error |
on: Apr 29th, 2007, 4:59pm |
| Started by Jashan | Post by Jashan |
Fixing the ANSI > UTF8 encoding on the text file solved the first error, but like you I encountered the memory-hang. After much experimentation, I narrowed it down to this rule as causing (at least the first) hang:
# Verner's Law
<UNST>/<VOST>/<VOSO>_(C)(C)(C)<STVO>
<UNFR>/<VOFR>/<VOSO>_(C)(C)(C)<STVO>
The problem is not the (C)'s, as it did the same thing with this variant:
# Verner's Law
<UNST>/<VOST>/<VOSO>_<STVO>
<UNFR>/<VOFR>/<VOSO>_<STVO>
After some further testing, it appears to be a problem with the declaration of STVO, which was:
# Vowels
V=[a:][e:][i:][o:][u:][y:][á:]aeiouyáéóí
# STRESSED VOWELS
STVO=[<V>']
Why this didn't work, I'm not sure, but when I deleted the STVO variable and changed the rule to this, the memory problem stopped. (Well, it became significantly less.)
# Verner's Law
<UNST>/<VOST>/<VOSO>_V'
<UNFR>/<VOFR>/<VOSO>_V'
Just FYI.
|
|
|
| 6 |
Versatile Sound Change Applier
/ Re: UTF8 error |
on: Apr 29th, 2007, 1:23pm |
| Started by Jashan | Post by MUBA |
Ok, found the problem, probably. It seems ruleset (which has some accented vowels) is not encoded in UTF-8, but in Latin-1.
I got rid of the error by loading the ruleset file in BabelPad (windows, but you seem to be on windows too so that's no problem) and saving it as UTF-8 again.
You could do that, or run VSCA with the -u switch, which disables the UTF-8 input/output layer - that also should get rid of the probem.
But alas, things are not as easy as they seem - this way you get rid of one problem, but on my machine a new one occurs: vsca silently runs for a while, until I get the dialog window with the white cross in a red circle, the illegal operation one. I'm not sure what's causing that, but considering how slow my computer gets when I use the files you gave, it might be a memory problem, which is not that weird, seeing as you're using lots of optionals and I have only 128mb of ram.
Anyway, tell me if this solves your problem. If you're getting any further problems, we can see what we can do about the memory thing (apart from buying additional ram, I mean  )
Yours,
MUBA
|
|
|
| 7 |
Versatile Sound Change Applier
/ Re: UTF8 error |
on: Apr 29th, 2007, 1:04pm |
| Started by Jashan | Post by MUBA |
Hmm, that is a rather weird error you're getting - it's not supposed to occur anymore. Let me investigate it.
|
|
|
| 8 |
Versatile Sound Change Applier
/ UTF8 error |
on: Apr 29th, 2007, 12:28pm |
| Started by Jashan | Post by Jashan |
I can't for the life of me figure out what I'm doing wrong. I've gone over the file at least 10 times looking for funny characters or things out of place and can't find it. Help?
I'm running Perl on Win XP Pro:
----------------------------------
C:\VSCA>perl -v
This is perl, v5.8.8 built for MSWin32-x86-multi-thread
(with 50 registered patches, see perl -V for more detail)
Copyright 1987-2006, Larry Wall
Binary build 820 [274739] provided by ActiveState http://www.ActiveState.com
Built Jan 23 2007 15:57:46
-----------------------------------------------
And receive the following error when trying to run a preliminary test of my sound changes:
----------------------------------------
C:\VSCA>perl vsca.pl -l latin.txt -r latingerman.txt -o output.txt -H
utf8 "\xE1" does not map to Unicode at vsca.pl line 984, <$fh> chunk 1.
Malformed UTF-8 character (unexpected end of string) in substitution (s///) at v
sca.pl line 85.
Malformed UTF-8 character (unexpected end of string) in substitution (s///) at v
sca.pl line 86.
Malformed UTF-8 character (unexpected non-continuation byte 0xf3, immediately af
ter start byte 0xe9) in substitution (s///) at vsca.pl line 86.
Malformed UTF-8 character (unexpected end of string) in substitution (s///) at v
sca.pl line 85.
Malformed UTF-8 character (unexpected end of string) in substitution (s///) at v
sca.pl line 86.
Terminating on signal SIGINT(2)
----------------------------------------------
Here's an example of the Lexicon. It's all straight Latin characters, with # and : as punctuation (and whitespace):
## PUER PUERI
puer
pueri:
puero:
puerum
puero
pueri:
puero:rum
pueri:s
puero:s
pueri:s
## VIR VIRI
vir
viri:
viro:
virum
viro
viri:
viro:rum
viri:s
viro:s
viri:s
----------------------------------------
And here's the rules file in all its temporary glory. Note that the board formatting messes up the comment lines, but they're all commented out at the beginning of the line.
# Germanic-Latin
# Variable declarations
# Unvoiced Stops
UNST=ptk
# Unvoiced Fricatives
UNFR=f[th]x
# Voiced Stops
VOST=bdg
# Voiced Fricatives
VOFR=v[dh]h
# Fricatives
FRIC=<VOFR><UNFR>
# Nasals
N=nm[ng]
# Short vowels 1
SHVO=aeiouyá
# Long vowels 1
LOVO=[<SHVO>:]
# Consonants
C=bdgptkfshmnrljv[ng]x[dh][th]q
# Vowels
V=<LOVO><SHVO>éóí
# STRESSED VOWELS
STVO=[<V>']
# Voiced Sounds (for Verner's Law)
VOSO=<V>bdgmnrljv[ng][dh]<STVO>
# Front vowels
FRONT=iejéí
# Adjustments for non-phonetic spelling
[th]/t/_
[ph]/p/_
c/k/_
[kh]/k/_
x/[ks]/_
[qu]/[kv]/_
i/j/_V
####################################
## THE ACTUAL RULES ##
####################################
# We don't like H's. Let's strengthen them. This is going to sound # like
Dutch after a while
h/x/_
# Grimm's Law
<UNST>/<UNFR>/_ UNLESS s_
<VOST>/<UNST>/_ UNLESS s_
# Verner's Law
<UNST>/<VOST>/<VOSO>_(C)(C)(C)V'
<UNFR>/<VOFR>/<VOSO>_(C)(C)(C)V'
# Loss of nasals before fricatives with compensatory lengthening of # the
preceding vowel
<SHVO>/<LOVO>/_N<FRIC>
N//_<FRIC>
# Loss of syllable-final M
m//_C OR _#
# Standard Latin vowel shifts from classical to proto-Romance
[ae]/é/_
[oe]/e/_
[au]/O/_
y/u/_
i/í/_ UNLESS _:
e/é/_ UNLESS _:
o/ó/_ UNLESS _:
u/o/_ UNLESS _:
://V_
# Short O/A merger
ó/a/_
# Reduction of /i/ and /e/ to /j/ when pre-vocal
i/j/_V
e/j/_V
é/j/_V
í/j/_V
# Loss of syllable-final T
t//_C OR _#
# Loss of intervocalic /v/
v//V_V
# Initial voiced obstruents devoice
<VOFR>/<UNFR>/#_
<VOST>/<UNST>/#_
# /v/ becomes /b/
v/b/_
# Unstressed vowels are lost in medial syllables
V//V(C)(C)(C)_(C)(C)(C)V UNLESS _'
# Nasal point-of-articulation assimilation
N/m/_b OR _p OR _m
N/[ng]/_g OR _k OR _x OR _[ng]
# Stress shifts to initial syllable (whoohoo! Germanic!)
'//_
V/V'/#(C)(C)(C)(C)_
# I-mutation
u/y/(C)(C)(C)(C)<FRONT>
o/y/(C)(C)(C)(C)<FRONT>
í/i/(C)(C)(C)(C)<FRONT>
e/i/(C)(C)(C)(C)<FRONT>
é/e/(C)(C)(C)(C)<FRONT>
a/á/(C)(C)(C)(C)<FRONT>
# Simplify geminates
[C+]/C/_
# Stop-stop and fric-fric pairs that share point-of-articulation # # simplify to
the voiceless counterpart of the first of the pair.
[bp]/p/_
[pb]/p/_
[dt]/t/_
[td]/t/_
[gk]/k/_
[kg]/k/_
[sth]/s/_
[ths]/[th]/_
[sdh]/s/_
[dhs]/[th]/_
[dhth]/[th]/_
[thdh]/[th]/_
[hx]/x/_
[xh]/x/_
# Beginning of a a voiceless stop > voicess fric shift (only k > x
# right now)
k/x/_ UNLESS _<FRIC>
# Devoicing of word-final obstruents
bdgv[dh]/ptkf[th]/_#
# Allophonic variation of /x/
x/h/_(C)<FRONT>
|
|
|
| 10 |
Bugs and feature requests
/ Re: [UTF-8] Perl warning: wide character in print |
on: Mar 8th, 2007, 6:52pm |
| Started by MUBA | Post by MUBA |
Solved all of that...
The "wide character in print" message doesn't show up anymore.
Different types of newlines are now supported - cross-platform, even.
And a very obscure problem with unicode and the substr() function got solved.
Which means...
NEW VERSION!
|
|
|
|
