lingdb / Sound-Comparisons

Exploring phonetic diversity across language families —
http://www.soundcomparisons.com
Other
13 stars 8 forks source link

New Malakula Languages (February 2017) #440

Open PaulHeggarty opened 7 years ago

PaulHeggarty commented 7 years ago

Aviva has sent in the new Malakula languages tree structure. Various tasks for Paul here, to update the database:

Presumably for the revised recordings the transcriptions will be altered on the existing Excel transcription pasting sheets. But for the 23 new varieties, we need:

And two processing tasks for Laura:

@AvivaShimelman, can you confirm I’ve got this right?

PaulHeggarty commented 7 years ago

Several of the new recordings are named the same as existing ones, but as version 2 of that one. E.g. Peskaros 2, Hokai 2, etc.. Some questions for @AvivaShimelman:

PaulHeggarty commented 7 years ago

@AvivaShimelman, can you also confirm an important question about the 23 new varieties to add: do they all just need inserting into the current nested structure, at levels that already exist for the existing languages, or are there cases that needed you to add new nodes higher up the nested structure (levels up to 5)?

PaulHeggarty commented 7 years ago

Another for @AvivaShimelman: Netimb now seems nested under Aulua, but currently it's called Netimb: Bangir, rather than Aulua: Netimb. It seems the same place and variety, so I've changed the lat-long as you indicated, but wanted to check this was all ok, and whether the display name should change.

PaulHeggarty commented 7 years ago

I’ve updated the latlongs as indicated. @AvivaShimelman, you might want to check they’re right now.

PaulHeggarty commented 7 years ago

By this stage, the spreadsheet that Aviva sent has become so outdated and messed about by multiple conversions between different forms of excel (PC vs. Mac) that it has become totally unworkable. There were many problems, including the nested structure being inconsistently ordered in many places, blank rows, and so on. I’ve had to radically recreate it. The new one is now on OwnCloud.

Please, please, please do not insert or delete any columns. Just complete any missing or incorrect data in there existing cells. Please also do not delete/insert any rows.

The colours Aviva used were all randomised in conversion, so I’ve marked as 'new' any entries that are either new, or altered.

I’ve given up on trying to ask for new data only to be entered in this pasting spreadsheet, and corrections to the Reference Data sheet. It will take me many hours to move to a new system, but it seems there's no solution. This 'pasting' sheet will replace the Reference Data sheet in due course, so please complete the speaker names and lat-longs in here, rather than in the reference data table. The reference data table will still be used for listing filenames for now, but nothing else.

There are many possible inconsistencies in naming varieties. You will see that there's a column for manually entering names to override the automatic system for generating them.

AvivaShimelman commented 7 years ago

Right.

AvivaShimelman commented 7 years ago

Yes. Not ideal. It was a result of the need to get non-inflected versions of the recordings. If there are only, say, three dialects of a language (common with languages with 80 speakers), and I already have three recordings on the site, if I have to do a new one, that new one will likely be of one of thoes same dialects. Two solutions:

  1. Just replace the old recording. But that loses a lot of information, and a lot of the old recordings are pretty for other reasons.
  2. Invent dialect names. No one will really recognize them, but they will be minimally meaningful. I did this for Neve'ei, for example, when I called the dialect "Coast."
  3. Live with the "2"s.

I suggest (2).

A.

On 2/13/17, Paul Heggarty notifications@github.com wrote:

Several of the new recordings are named the same as existing ones, but as version 2 of that one. E.g. Peskaros 2, Hokai 2, etc.. Some questions for @AvivaShimelman:

  • Are these just multiple speakers of exactly the same variety?
  • Should they both be up on the website? (Pretty messy to display them, currently.)
  • Should version 2 replace version 1?
  • What's the logic for having 2 speakers? For Sound Comparisons purposes, I generally try to avoid this.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-279348454

-- Aviva Shimelman, PhD

AvivaShimelman commented 7 years ago

Only one case with a new higher (level 5) node: Larevat. It is the unique entry (numbered 3) under West (level 3 - (2)) and North (level 2 - (1)

A.

On 2/13/17, Paul Heggarty notifications@github.com wrote:

@AvivaShimelman, can you also confirm an important question about the 23 new varieties to add: do they all just need inserting into the current nested structure, at levels that already exist for the existing languages, or are there cases that needed you to add new nodes higher up the nested structure (levels up to 5)?

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-279348871

-- Aviva Shimelman, PhD

AvivaShimelman commented 7 years ago

After an hour (no joke!) with the speaker getting the former boundaries of the language right, I couldn't be more confident of the latlongs I entered for it. Calling it Bangir would have been a mistake. Bottom line, call it Aulua: Netimb.

Longer explanation - Netimb:Aulua makes sense in a way. Speakers refer to Netimb as a "language." It has 7 speakers, all of whom have been living in different AUlua-speaking communities all their lives. As recorded now, it is 96% cognate with Aulua, although, of course, it was probably formerly much more distinct.

A.

On 2/13/17, Paul Heggarty notifications@github.com wrote:

Another for @AvivaShimelman: Netimb now seems nested under Aulua, but currently it's called Netimb: Bangir, rather than Aulua: Netimb. It seems the same place and variety, so I've changed the lat-long as you indicated, but wanted to check this was all ok, and whether the display name should change.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-279351751

-- Aviva Shimelman, PhD

PaulHeggarty commented 7 years ago

I have now just finished a significant update to the excel spreadsheet, updated now on OwnCloud, please now go ahead and make corrections to that if necessary.
I’ve also attempted to put all new languages on the website. Obviously, there are now some problems showing up there, but I wanted to do this so that both of you, @AvivaShimelman and @LauraWae, can see where any problems are. Once you’ve checked it over we can also temporarily switch off the problematic varieties until they are transcribed and uploaded, if you wish.

PaulHeggarty commented 7 years ago

Please let me know if there's anything in the old reference data sheet that is still missing from the new for pasting sheet.

AvivaShimelman commented 7 years ago

Got it.

Some answers, some questions:

wrt Languages without unique codes.

On this sheet, it does look like there are three cases where level 5 numbers with the same level 4 and 3 numbers are doubled.

One by one.

It looks like Wala Preng and Wala Small Island (both 7-5-49) have the same number. Preng is a specific place on the small island. There is no recording "Wala Small Island" on the site. I'm guessing that at one point we replaced the one with the other. Indeed, on the Reference Sheet for, for example, 2016 5 11 (long after we would have stopped playing with the UWRAs), only Preng appears.

On this sheet, Caroline Bay and Elu do indeed have the same number, (7-1-85). On my earlier pasting sheets, though, these are listed as 85 and 86, respectively, so I don't know what accounts for the slide. I wouldn't have touched them.

The two Bislamas do have the same number. Looks like it's been that way from the beginning. But on the reference sheets the two are not indexed with numbers that read off the sheets, but rather with 487000007 and 8.

I see no case where there is no number at all on my last sheet.

So, yes, issues with multiple conversions, multiple hands, multiple generations of the site, multiple changes to practice ...

A

On 2/13/17, Paul Heggarty notifications@github.com wrote:

By this stage, the spreadsheet that Aviva sent has become so outdated and messed about by multiple conversions between different forms of excel (PC vs. Mac) that it has become totally unworkable. There were many problems, including the nested structure being inconsistently ordered in many places, blank rows, and so on. I’ve had to radically recreate it. The new one is now on OwnCloud.

Please, please, please do not insert or delete any columns. Just complete any missing or incorrect data in there existing cells. Please also do not delete/insert any rows.

  • [ ] Several cases of duplicated level 6 numbers (00-99). This is not possible, theyou need to be at least one digit different. Often done when there's a speaker 1 and speaker 2. But they still need minimally different numbers at level 6.
  • [ ] Some cases of no number at all at level 6.

The colours Aviva used were all randomised in conversion, so I’ve marked as 'new' any entries that are either new, or altered.

I’ve given up on trying to ask for new data only to be entered in this pasting spreadsheet, and corrections to the Reference Data sheet. It will take me many hours to move to a new system, but it seems there's no solution. This 'pasting' sheet will replace the Reference Data sheet in due course, so please complete the speaker names and lat-longs in here, rather than in the reference data table. The reference data table will still be used for listing filenames for now, but nothing else.

There are many possible inconsistencies in naming varieties. You will see that there's a column for manually entering names to override the automatic system for generating them.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-279429167

-- Aviva Shimelman, PhD

AvivaShimelman commented 7 years ago

On it.

On 2/13/17, Paul Heggarty notifications@github.com wrote:

I have now just finished a significant update to the excel spreadsheet, updated now on OwnCloud, please now go ahead and make corrections to that if necessary. I’ve also attempted to put all new languages on the website. Obviously, there are now some problems showing up there, but I wanted to do this so that both of you, @AvivaShimelman and @LauraWae, can see where any problems are. Once you’ve checked it over we can also temporarily switch off the problematic varieties until they are transcribed and uploaded, if you wish.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-279523609

-- Aviva Shimelman, PhD

PaulHeggarty commented 7 years ago
  1. So I'll delete Wala Small Island and leave only Wala Preng, OK?
  2. I'll set Elu to 86 then.
  3. Bislama is a special case which I’ve tweaked individually. Not a problem now.
PaulHeggarty commented 7 years ago

Yes, the explanation lies in all of those multiples you mention. That's just how it is. No worries.

AvivaShimelman commented 7 years ago

SOunds right.

On 2/13/17, Paul Heggarty notifications@github.com wrote:

  1. So I'll delete Wala Small Island and leave only Wala Preng, OK?
  2. I'll set Elu to 86 then.
  3. Bislama is a special case which I’ve tweaked individually. Not a problem now.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-279528292

-- Aviva Shimelman, PhD

AvivaShimelman commented 7 years ago

Will do. Any big mistakes will be obvious right away on the site, of course.

On 2/13/17, Paul Heggarty notifications@github.com wrote:

I’ve updated the latlongs as indicated. @AvivaShimelman, you might want to check they’re right now.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-279352404

-- Aviva Shimelman, PhD

AvivaShimelman commented 7 years ago

Short answer to immediate question:

Yes. Speaker names had been filled in. (They now appear just as notes). And there seem to be a bunch of ISO and glotocodes missing even for languages for which they exist. Also missing for the same languages: L and L. All affected languages seem to be dialects of languages that appeared with single representatives in the first demo site (December 2015).

Answers to questions about the languages as now showing on the website:

New Vao (1-5-5-75) not appearing on map but listed among languages. One old Vao (Peteri (1-5-5-76)) missing both on list and map

Missing new languages: Natanggan Veretemate Nasvang Pimilis Unua Benuamor (all, not surprisingly, without level 5 codes on new sheet)

The "2"s did create issues. It looks like the sound is playing from the old "1" when you mouse over the "2"; moreover, these do not have "soons" as do the others, but play arrows. Affected: Avok (3-9-7-64 (the one with a doubled level 5)), Uliveo (Maskelynes) (3-9-8-56), Nisvai (3-9-5-78) I will give these different names.

LL Maraxus Mae seems off. Will check. The OLD Nisvai seems to have moved. Will check

Also: Naha'ai and Naxa'ai (2-9-5) are still appearing (my fault!) as Na'ahai and Na'axai (common variants in speaker production, btw)

Old recordings now missing? Uripiv: Tautu (1-7-5-87)

RE: The "double" Bangaasak on the pasting sheet. It is a dialect with a single speaker, so there's no way there could be multiple recordings. It must just have gotten copied at some point.

Recordings slated to go up a couple of months ago: Three V'enen Tautes

LAURA: My fault. The word playing for 'woman' (and the transcription appearing with it) for Naman is, in fact, the word for 'what' (which also appears again in the right place, too). I've checked the other entries for the language and all seem fine.

Will fine-tooth-comb it again after a shake-out.

A.

On 2/13/17, Paul Heggarty notifications@github.com wrote:

Please let me know if there's anything in the old reference data sheet that is still missing from the new for pasting sheet.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-279523949

-- Aviva Shimelman, PhD

AvivaShimelman commented 7 years ago

Excellent! The two-sheet system just opened up the door to too many inconsistencies, in my view. We're working with two different sets of sheets with multiple iterations, so it's hard, even with the dates, to know what is current, i.e., what trumps what.

It looks like the new "00" sheet (love the colors!) imported the L and Ls off of the last (2017 02 12) "for pasting" sheet. The problem is these don't always correspond to the L and Ls on the last version of the reference sheet (probably because these didn't get "back-updated" when the newer sheets were). Examples of things I can identify for now:

The second Bislama recording is listed on the "00" sheet with the same L and L as the first one is, but in the last reference sheet (2016 11 11), it is given as -16.10302 167.4168. The Espiegel's Bay recordings also have different L and Ls, as do TIrax Bakru, Uripiv Potnambe, Uripiv D'uripiv, Wowo Lesmarlas, Maraxus, Port Sandwich Lamap 2, Avava Tisvel.

I cannot check for consistencies with the others because they are missing.

I can figure out which is the most recent and correct it if it needs to be corrected. The question is -- just to be absolutely clear -- where am I supposed to enter any corrections and how I am I supposed to signal that these are corrections?

Other issues:

Missing recordings on 00: 2-9 (West South) There is one language, Na[h,x]a'ai, with two dialects: Naha'ai and Naxa'ai. We have two sub-dialects of each of these, Toman and Mbatmbang and Malfaxal and Mbatvanui, respectively, These all are all listed as Na'ahai on the RS (which I may or may not have done, as there are reasons for going both routes), but, whatever the case with that, only two sub-dialcets, one representative of each, seem to make it to the 00 sheet, Mbatmbang and Malfaxal.

Also missing, I think: 3-9-8 (Uliveo) dialects Lurtes and Pelongk (sometimes it's hard to tell because dialcets are not always grouped together on the RSs)

Also missing (on both): 3 added dialects for V'enen Taute: Benenavet, Benwara, and Tenmaru (I think) (two of these are sitting in the "to clean up" folder, which is good evidence that we did add them at one time.

Will keep inspecting.

Hope all are long asleep by now.

A.

On 2/13/17, Paul Heggarty notifications@github.com wrote:

By this stage, the spreadsheet that Aviva sent has become so outdated and messed about by multiple conversions between different forms of excel (PC vs. Mac) that it has become totally unworkable. There were many problems, including the nested structure being inconsistently ordered in many places, blank rows, and so on. I’ve had to radically recreate it. The new one is now on OwnCloud.

Please, please, please do not insert or delete any columns. Just complete any missing or incorrect data in there existing cells. Please also do not delete/insert any rows.

  • [ ] Several cases of duplicated level 6 numbers (00-99). This is not possible, theyou need to be at least one digit different. Often done when there's a speaker 1 and speaker 2. But they still need minimally different numbers at level 6.
  • [ ] Some cases of no number at all at level 6.

The colours Aviva used were all randomised in conversion, so I’ve marked as 'new' any entries that are either new, or altered.

I’ve given up on trying to ask for new data only to be entered in this pasting spreadsheet, and corrections to the Reference Data sheet. It will take me many hours to move to a new system, but it seems there's no solution. This 'pasting' sheet will replace the Reference Data sheet in due course, so please complete the speaker names and lat-longs in here, rather than in the reference data table. The reference data table will still be used for listing filenames for now, but nothing else.

There are many possible inconsistencies in naming varieties. You will see that there's a column for manually entering names to override the automatic system for generating them.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-279429167

-- Aviva Shimelman, PhD

LauraWae commented 7 years ago

Hi both, I have carefully read your conversation.

However, concerning V'enen Taute: Benenavet and V'enen Taute: Benwara:

Also missing (on both): 3 added dialects for V'enen Taute: Benenavet, Benwara, and Tenmaru (I think) (two of these are sitting in the "to clean up" folder, which is good evidence that we did add them at one time.

No, we have not added them yet. When I saw them in October last year, when we were uploading the previous bunch of files, I looked up for an entry in the Reference Sheet and there was none. So, I decided to leave them in "Don't worry about just yet" for the next round.

So, yes, we need still the information for those records. Thanks a lot!

LauraWae commented 7 years ago

Concerning this:

LAURA: My fault. The word playing for 'woman' (and the transcription appearing with it) for Naman is, in fact, the word for 'what' (which also appears again in the right place, too). I've checked the other entries for the language and all seem fine.

That means, it is highly recommendable to exchange or quit that one, right? Will do. Keep track on #372.

AvivaShimelman commented 7 years ago

Sorry. I remember that this was a conversation we had just before I left (and you, too) on our last field trips. I think we decided that that swath just looked too bare and, even if it were just for aesthetic reasons, it made sense to include a couple more V'enen Tautes. It's altogether possible that I never amended the sheet. I'll include those with the new batch as soon as I have a "final word" on our present process for adding languages and amending information for old entries.

A.

On 2/14/17, LauraWae notifications@github.com wrote:

Hi both, I have carefully read your conversation and see that there is no a lot I can contribute.

However, concerning V'enen Taute: Benenavet and V'enen Taute: Benwara:

Also missing (on both): 3 added dialects for V'enen Taute: Benenavet, Benwara, and Tenmaru (I think) (two of these are sitting in the "to clean up" folder, which is good evidence that we did add them at one time.

No, we have not added them yet. When I saw them in October last year, when we were uploading the previous bunch of files, I looked up for an entry in the Reference Sheet and there was none. So, I decided to leave them in "Don't worry about just yet" for the next round.

So, we still need all information for those.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-279701872

-- Aviva Shimelman, PhD

AvivaShimelman commented 7 years ago

Right. Just kill that word please.

I'm guessing now that the consultant, a bit hard of hearing, heard "wanem" ('what') when I said "woman." I checked entries against Crowley's lexicon, but, as that language, Naman, is one of those two-speakers-left languages, I didn't have a chance to check it against the production of another speaker. I probably should have noticed that the word got repeated, but with 20,000 items ... Thanks.

A.

On 2/14/17, LauraWae notifications@github.com wrote:

Concerning this:

LAURA: My fault. The word playing for 'woman' (and the transcription appearing with it) for Naman is, in fact, the word for 'what' (which also appears again in the right place, too). I've checked the other entries for the language and all seem fine.

That means, it is highly recommendable to exchange or quit that one, right? Will do.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-279735973

-- Aviva Shimelman, PhD

AvivaShimelman commented 7 years ago

I've uploaded a spreadsheet, "2017 03 13 For Pasting in to the CURRENT Malakula Tree Structure for Pasting.xls" to owncloud. It is attached here as well as a .xlsx. It is built on the "00 CURRENT ..." sheet that Paul uploaded two days ago. As I didn't want to risk changing anything with changes in software, formats, etc., I included only the new languages (Two Larevats and one Ngava) along with those that seemed to have gotten dropped at some point. I've dropped breadcrumbs to leave a trail back to the first node, so it should be clear where these paste into the existing sheet. All "returning" languages have the exact same information as before. In particular, I have made no changes to any part of their index numbers. All the languages except the Venen Tautes seem to have made their auto-calculations. I can paste the elements of this sheet into that CURRENT sheet from two days ago as long as no one else is nervous about that. 2017 03 13 For Pasting in to the CURRENT Malakula Tree Structure for Pasting.xlsx

PaulHeggarty commented 7 years ago

I have done new versions of both main sheets in the OwnCloud x Languages Reference folder. I’ve also updated the website to the list currently showing in the Reference file.
This includes creating the new languages Aviva requested.
Can you look at the sheets, and the website, to see how much this fixes, and what remains not working.

AvivaShimelman commented 7 years ago

On it.

On 3/14/17, Paul Heggarty notifications@github.com wrote:

I have done new versions of both main sheets in the OwnCloud x Languages Reference folder. I’ve also updated the website to the list currently showing in the Reference file. This includes creating the new languages Aviva requested. Can you look at the sheets, and the website, to see how much this fixes, and what remains not working.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-286422922

-- Aviva Shimelman, PhD

LauraWae commented 7 years ago

Hi both,

Please note that this sheet does not include the Recordings' Original File Names of the latest batch of languages. Ideally, they would be also included here.

As I it is a good time to progress with the uploading now, could you please tell me @AvivaShimelman, in which of the four excel sheets onownCloud\SndComp\Malakula\z Languages Reference, I can find the Original Files names AND their corresponding SoundComparison Names for the new languages (Oce_Van_Mal...) in a row?

I hope the Original file names can be added later then to the Reference Sheet.

Thanks to both!

AvivaShimelman commented 7 years ago

The sheet I uploaded on Saturday does indeed have the file names. Please consult 2017 0310 by Aviva .." in the old versions subfolder of z language reference.

A.

On 3/14/17, LauraWae notifications@github.com wrote:

Hi both,

Please note that this sheet does not include the Recordings' Original File Names of the latest batch of languages. Ideally, they would be also included here.

As I it is a good time to progress with the uploading now, could you please tell me @AvivaShimelman, in which of the four excel sheets on ownCloud\SndComp\Malakula\z Languages Reference, I can find the Original Files names AND their corresponding SoundComparison Names for the new languages (Oce_Van_Mal...) in a row?

I hope the Original file names can be added later then to the Reference Sheet.

Thanks to both!

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-286435418

-- Aviva Shimelman, PhD

LauraWae commented 7 years ago

Where?

image

PaulHeggarty commented 7 years ago

I can see filenames for the recently added languages in: C:\Paul\« Cloud\OwnCloud\SndComp\Malakula\z Languages Reference\OLD Versions\OLD Nested Tree Structure\2017 03 08b Malakula Tree Structure for Pasting.xlsm …. so they can be copied and pasted in from there.

AvivaShimelman commented 7 years ago

Column 15.

AvivaShimelman commented 7 years ago

It looks like that is not in fact "by Aviva". Maybe it was edited when it was renamed? In any case, the file names for the new languages can be found on the new for pasting sheet I uploaded two days ago, dated the 13th and included in my post above.

AvivaShimelman commented 7 years ago

Here's another copy of it. 2017 03 11 CURRENT Malakula Tree Structure for Pasting (1) (by Aviva, in fact a reference sheet).xlsx

AvivaShimelman commented 7 years ago

To respond first just about the sheets. Some information has been lost on both, but not the same information. Both do contain what I believe to be a full census. No MIAs no double personalities. Yuppee! The reference sheet has lost file name and particpant name information -- as well as L and L -- for the languages added on the "for pasting for pasting" sheet. The For Pasting sheet, in contrast, suffered a bit more, losing participant names for all but the new files, as well as file names and L and L for all the first generation recordings, and, I think, some of the manual over-ride titles.

On 3/14/17, Paul Heggarty notifications@github.com wrote:

I have done new versions of both main sheets in the OwnCloud x Languages Reference folder. I’ve also updated the website to the list currently showing in the Reference file. This includes creating the new languages Aviva requested. Can you look at the sheets, and the website, to see how much this fixes, and what remains not working.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-286422922

-- Aviva Shimelman, PhD

PaulHeggarty commented 7 years ago

Could you both give me 20 minutes and I’ll try to fix it all once and for all. Don’t bother doing anything else yourselves in the meantime!

AvivaShimelman commented 7 years ago

SC 2017 03 14 check:

Systematic: All languages added in the past month are showing up “soon” All the languages that got revised are showing up without sound All of the recordings (all from the south, all from the fall field trip batch) that had gone missing from the sheets are appearing with sound but no transcriptions (i.e., with play triangles).

Quasi systematic? Neverver Limap is new, but Neverver Sakan and Mindu are old. We may have turned them off in November, pending verification. They can probably come back on again. I'll triple-verify, though. Ninde Lawa isn't playing (maybe a November “turn off”). It, too, can come back from the abyss.

Idiosyncratic issues: Maraxus: Mae has lost sound.LAURA: this was doubled with Maraxus: Maraxus. Did the sound and transcription get assigned to this language name? I'm not sure why Naati changed from Naati:Windua to Naati:Naati or why it wouldn't be playing. No sound for Nahavaq Caroline Bay or Naha'ai Toman Island (both older recordings). The latter had gone MIA for a while. Nisvai: Blacksands has lost transcription Aulua: Loxse Lanfitfit has lost sound Njav: Gonwar has lost sound

In other news: Wala Preng is floating a bit.

A.

On 3/14/17, Paul Heggarty notifications@github.com wrote:

I have done new versions of both main sheets in the OwnCloud x Languages Reference folder. I’ve also updated the website to the list currently showing in the Reference file. This includes creating the new languages Aviva requested. Can you look at the sheets, and the website, to see how much this fixes, and what remains not working.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-286422922

-- Aviva Shimelman, PhD

PaulHeggarty commented 7 years ago

I have just updated as much as I can, and there are very few gaps left now. I’ve left them in orange in 00 CURRENT Malakula Language Reference Data on OwnCloud. Aviva, could you check and fill in the last gaps, and let me know when you’re done so I can upload them tonight. Sorry if this is data you have entered before, but there really is very little now, and I hope this will be done once and for all.

AvivaShimelman commented 7 years ago

OK

On 3/14/17, Paul Heggarty notifications@github.com wrote:

Could you both give me 20 minutes and I’ll try to fix it all once and for all. Don’t bother doing anything else yourselves in the meantime!

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-286486650

-- Aviva Shimelman, PhD

AvivaShimelman commented 7 years ago

On it.

On 3/14/17, Paul Heggarty notifications@github.com wrote:

I have just updated as much as I can, and there are very few gaps left now. I’ve left them in orange in 00 CURRENT Malakula Language Reference Data on OwnCloud. Aviva, could you check and fill in the last gaps, and let me know when you’re done so I can upload them tonight. Sorry if this is data you have entered before, but there really is very little now, and I hope this will be done once and for all.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-286495968

-- Aviva Shimelman, PhD

AvivaShimelman commented 7 years ago

Joy! I just uploaded to OC my corrected version of this evening's reference sheet. All blanks are genuine blanks (i.e., languages without ISOs genuinely do not have assigned ISOs). I have made about a half-dozen unsolicited changes that I have signaled in burnt orange. Go, Team Malakula!

On 3/14/17, Paul Heggarty notifications@github.com wrote:

I have just updated as much as I can, and there are very few gaps left now. I’ve left them in orange in 00 CURRENT Malakula Language Reference Data on OwnCloud. Aviva, could you check and fill in the last gaps, and let me know when you’re done so I can upload them tonight. Sorry if this is data you have entered before, but there really is very little now, and I hope this will be done once and for all.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-286495968

-- Aviva Shimelman, PhD

PaulHeggarty commented 7 years ago

I’ve filled in all of those in the database, and uploaded the new version to the website. Please let me know if you notice anything that still hasn't worked.

AvivaShimelman commented 7 years ago

Paul? It's loading exactly as before in both Chrome and safari.

Another issue: It loads zoomed out to the whole world.

A.

On 3/14/17, Paul Heggarty notifications@github.com wrote:

I’ve fillen in all of those in the database, and uploaded the new version to the website. Please let me know if you notice anything that still hasn't worked.

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-286580119

-- Aviva Shimelman, PhD

PaulHeggarty commented 7 years ago

For me it's working: here's one of the changes to one of the Siviti names, showing correctly:
screenshot - 2017-03-15 08_11_57 The world zoom problem is one there's an issue for. The difficulty is that it's sporadic. Mine has just loaded fine!
Can you do a forced refresh, Ctrl-R or similar, and see if that fixes anything?

AvivaShimelman commented 7 years ago

I've not only refreshed, I've cleared cache, opened different browsers and used two different computers. I think we're not talking about the same issue. The languages are indeed listed. I thought you were talking about the fact that there are some that aren't playing or that transcriptions aren't showing up.

A.

On 3/15/17, Paul Heggarty notifications@github.com wrote:

For me it's working: here's one of the changes to one of the Siviti names, showing correctly: screenshot - 2017-03-15
08_11_57 The world zoom problem is one there's an issue for. The difficulty is that it's sporadic. Mine has just loaded fine! Can you do a forced refresh, Ctrl-R or similar, and see if that fixes anything?

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-286659769

-- Aviva Shimelman, PhD

PaulHeggarty commented 7 years ago

!! If only it were so simple! All I did was upload the correct data to the website, so that the right languages appear with the right names, and speakers if there are photos. That cannot magically rename sound files, nor copy, paste and load up transcriptions and sound files. All that needs Laura to have the correct cross-referencing data from the Languages Reference sheet, plus lots more work for her. My uploading was stage 3 of 50 or so... ;)

AvivaShimelman commented 7 years ago

"TLC"

On 3/15/17, Paul Heggarty notifications@github.com wrote:

!! If only it were so simple! All I did was upload the correct data to the website, so that the right languages appear with the right names, and speakers if there are photos. That cannot magically rename sound files, nor copy, paste and load up transcriptions and sound files. All that needs Laura to have the correct cross-referencing data from the Languages Reference sheet, plus lots more work for her. My uploading was stage 3 of 50 or so... ;)

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-286777479

-- Aviva Shimelman, PhD

PaulHeggarty commented 7 years ago

@AvivaShimelman , could you confirm that the correct name is Uliveo: Lurtes with an L, not Uliveo: Hurtes with an H, as it once appeared in an excel sheet. I presume a typo.

PaulHeggarty commented 7 years ago

@AvivaShimelman , could you also confirm that the correct spelling is Natangan with one g, rather than Natanggan with two gs. Thanks.

AvivaShimelman commented 7 years ago

L.

On Sun, Apr 2, 2017 at 3:28 PM, Paul Heggarty notifications@github.com wrote:

@AvivaShimelman https://github.com/AvivaShimelman , could you confirm that the correct name is Uliveo: Lurtes with an L, not Uliveo: Hurtes with an H, as it once appeared in an excel sheet. I presume a typo.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-291008857, or mute the thread https://github.com/notifications/unsubscribe-auth/ARGKM6BVtHoo_Xkij4IfMMzn2NYGzj4Hks5rr_bXgaJpZM4L_BOS .

-- Aviva Shimelman, PhD

AvivaShimelman commented 7 years ago

Could go either way. It's a pre-nasalized [g]. Your call. I lean, now to going with 1.

On Sun, Apr 2, 2017 at 3:30 PM, Paul Heggarty notifications@github.com wrote:

@AvivaShimelman https://github.com/AvivaShimelman , could you also confirm that the correct spelling is Natangan with one g, rather than Natanggan with two gs. Thanks.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/lingdb/Sound-Comparisons/issues/440#issuecomment-291008980, or mute the thread https://github.com/notifications/unsubscribe-auth/ARGKM_lq6b8VFeQaPPxCX-1hiBzAAuvVks5rr_dAgaJpZM4L_BOS .

-- Aviva Shimelman, PhD