Page MenuHomePhabricator

key mapping issues in Narayam Gujarati transliteration
Closed, ResolvedPublic

Description

During the intensive use, we have identified a few incorrectly mapped key behaviours, will appreciate if these are fixed ASAP:

_ has no mapping, should be mapped to _

~ is mapped to ્, which should be ~

ઁ is typed on key strokes Mm, which causes issues, as this is a common sequence for many words, Mm should result in ંમ્ (because of this behavious 'hiMmat' results in હિઁઅત, which should be હિંમત). so in nutshell 2 things here,

Mm should be ંમ્   and
MM should result be mapped to ઁ

ai is mapped to એ, which should be ઐ (એ is simply e) when typed with consonents, it displays correctly (કૈ, મૈ, etc) but when typed as wovel, there is issue.

oM is mapped to ૐ, which should be ઓં
ૐ should be mapped to OM

J needs to be mapped to ઼


Version: unspecified
Severity: normal

Details

Reference
bz35931

Event Timeline

bzimport raised the priority of this task from to Medium.Nov 22 2014, 12:16 AM
bzimport set Reference to bz35931.

Can this be looked at priority now? it's been a month since this bug was created...

Sorry for the delay and thanks for the heads up.

This is easy to fix, but I need some clarifications:

  1. '_': It is currently mapped to the Zero-Width Non-Joiner. Do you need the Zero-Width Non-Joiner character?
  1. '~': Do you need to map any other character to the virama?
  1. ઁ is already mapped to M^, for example aM^. Do you need MM too?
  1. Can you please give an example word with ઁ? It will be useful for testing.
  1. ` (backtick, to the left of digit 1) is currently mapped to ઼. Do you need 'J', too?

I may have a couple of questions later.

(In reply to comment #2)

  1. '_': It is currently mapped to the Zero-Width Non-Joiner. Do you need the

Zero-Width Non-Joiner character?

Yes, if you could assign ZWNJ to ` it would be great, release _ at it might be useful as _ only...

  1. '~': Do you need to map any other character to the virama?

Not really, as all the characters by default are khandit (with virama)

  1. ઁ is already mapped to M^, for example aM^. Do you need MM too?

No, in that case leave it as M^, the only thing to do with ઁ is to delink it from Mm

  1. Can you please give an example word with ઁ? It will be useful for testing.

very rarely used, can't think of but you can consider હિઁમત or અઁબર

  1. ` (backtick, to the left of digit 1) is currently mapped to ઼. Do you need

'J', too?

Releas ` and map ઼ to J or K

I may have a couple of questions later.

Most welcome to ask as many you wish...

Thanks for taking this in your hands, very much appreciated.

I made most of the changes, but

  1. For better testing I also need words which have એ, ઐ and કૈ, and also words with ઍ (candra e), as well as any variations of these letters.
  1. This proposal has some changes that make this transliteration scheme quite different from other Indic transliteration schemes. For example, most schemes use _ for ZWNJ and ` for Nukta. It is OK if the changes are useful for this language and similarity to other Indic input methods is not important to you, but just to be on the safe side it would be useful to get some more opinions about these changes.

(In reply to comment #4)

  1. This proposal has some changes that make this transliteration scheme quite

different from other Indic transliteration schemes. For example, most schemes
use _ for ZWNJ and ` for Nukta. It is OK if the changes are useful for this
language and similarity to other Indic input methods is not important to you,
but just to be on the safe side it would be useful to get some more opinions
about these changes.

Most of the changes that I suggested were already part of Gujarati transliteration, see local translit.js http://gu.wikipedia.org/wiki/%E0%AA%B8%E0%AA%AD%E0%AB%8D%E0%AA%AF:Dsvyas/Mediawiki:Translit.js and hence, this is the layout all the Gujarati wikipedians are used to with. In fact Narayam has caused the issues and disabled people to use the keyboard as they practiced so far... there is no standard transliteration layout in Gujarati, google and MS like major playes also differ on few things... other websites e.g. www.gujaratilexicon.com has slightly different layout then these, gu.wiki had slightly different but more user friendly version... fundamentally all are same, when it comes to sounds/phonetics, but special characters and signs are different everywhere...

Available for testing at the following link:

http://sandbox.translatewiki.net/wiki/User:Amire80/narayam-gu

Hurry up: We use that website for testing and it changes frequently :)

ૐ શ્રી ૧। (OM) shree (1K)
અમીરભાઈ, તમે ઐતિહાસીક ઝડ઼પથી (zaDJapathI) અને અદ્‌ભુત (ad`bhuta) રીતે આ નુક્ત અને ચઁદ્રબિંદીનું (caM^drabiMdI) કામ કરી આપ્યું, આશા રાખું કે હંમેશા (haMmeshaa) તમે અમને ગુજરાતીઓને આમ જ મદદ કરતા રહો.

Very well done Amir, thanks a million. Above message is the test, and I confirm all working perfectly well. I express my Special Thanks to you Amir, for providing me some extra support by offering help on the characters that I had not mentioned in my bug, e.g. danda (।) and making me aware of ZWNJ.

I will update the transliteration documentation once these changes are committed.

A sidenote: I acceidentally found that character દ્ભ (dbha) is missing in lohit fonts. Saying that, when lohit is selected dbh renders દ્‌ભ (just like the one with ZWNJ - dbha) and when switched to default Gujarati fonts, it becomes દ્ભ. But that has nothing to do with this bug. If it is missing in font family only, forget it, or else let me know, I'll file a new bug.

and yes, forgot to mention that tides are working perfectly well too (I missed to confirm that in above). I Sign off.

The change was approved and will be deployed some time soon.

(In reply to comment #10)

The change was approved and will be deployed some time soon.

When are these changes going to be deployed? How long does it generally take? Hope someone remembers to deploy this....

Except for high priority/emergency deployments it goes like this:

  • Change is submitted to Gerrit for review
  • Either directly or indirectly ends up in the master branch of the project
  • Next wmf-deployment cycle all extensions will be updated to the latest master

Right now we're on 1.20wmf3, the next cycle is 1.20wmf4 which happens to start today[1]. A cycle takes about about 2-4 weeks.

If this change was committed, review and merged into master before last weekend, it will be in 1.20wmf4 and rolled out over all wikis during the next 2 weeks.

[1] https://www.mediawiki.org/wiki/MediaWiki_1.20/Roadmap