Hi Akshat and the NBGP team, Please find below the responses inline below: On Mon, Aug 7, 2017 at 2:01 PM, Akshat Joshi <akshatj@cdac.in> wrote:

Hello Dr. Bal and our Nepal team,

Thanks for putting together the document for Nepali and Newar languages for Devanagari script.

Following are my observations:

1. Code point repertoire for the Devanagari Nepali and Newar languages

- As I see, the currently shared Devanagari LGR takes into account all the characters required by Nepali and Newar. Please let me know if there is some discrepancy related to the same. Also, it would be great if you could cite some additional references in the last column "References" which depict use of the individual characters in everyday use. They could be different references for different characters.

As far as the references for the keyboard layouts are concerned, different keyboard layouts have evolved lately but we basically use the "Traditional Keyboard Layout" and the "Romanized Keyboard Layout" for typing Unicode Nepali characters. Details on these layouts can be found in the link below: http://ltk.org.np/keyboard_layouts.php

- Also, as required, the 0931 (Ra Nukta) has been contextually permitted to form eyelash reph only.

Great, thanks.

2. Composite characters – Confusingly similar shapes

- The similar looking cases which are mere confusions on part of the user may not form part of the Devanagari variants. As per the LGR procedure, these cases are subject to "String Similarity Assessment" panel. As discussed in the Kathmandu meeting, I can definitely put them in Appendix with a reference for the string similarity panel to take them into account as an official recommendation from the NBGP.

OK.

Having said the above, I myself have contradicted to the above by including some of the "confusingly similar" cases as a part of variant recommendations for LGR. These are the cases pertaining to Santhali combinations where Nukta is expected to come with certain Vowels and Vowel signs. These are Unique cases because the "non-Santhali user-base of Devanagari" (which is major part of it) may not at all imagine presence of Nukta at those locations. Such instances may thus be construed by them as Stylistic variants/rendering problems thereby not making them sound an alarm. The point being, these are not mere visual similarity cases as they involve a congnitive lapse. This makes them worth being explicitly cited as variant.

I had a general query here. To which category would these variants be placed under - "allocatable" or "blocked" ? It it is under "allocatable", that should not be much of an issue from a language perspective but if it is assigned as "blocked", that could be blocking for genuine domain name entries of a particular language.

- Regarding similarity based on fonts:

I would request to refer to our discussion as per mail on 28th July '17 on the topic.

Fine, thanks.

- Regarding similarity between र + ् + इ and ई:

This is already being barred by our context rules which are based on earlier work done by C-DAC for .bharat domain names.

Great, thanks.

3. Homophonic variants

- As rightly pointed out in your document, these rules may not uniformly apply across the board to all the languages using Devanagari. Most of the suggestions under this section fall under the spelling norms which is not what we are aiming through LGR creation. An example for the same in English is e.g. No three consonants can come together to form a meaningful word, however fli*ckr* is still a domain name widely accepted and used by the Internet Community. In the same spirit, we will restrict ourselves from going in the "spelling norms". Also, not all spelling norms are algorithmically predictable and vary a great deal across the community.

As far as variant aspect of such words is concerned, there are two things about it.

- As per classical approach of domain name system, such cases are not treated as variants as their appearance is completely different. e.g. color vs colour.

- Also, even though it may appear that such cases can be algorithmically predicted going by the varga classification in Brahmi, across linguistic communities, these cases differ. The varga classification and it's last nasal consonant is perfect system in itself for predicting nasalization and conjuct behaviors in words, however it is not how it has come down into popular usage across the communities. The point being, it cannot be algorithmically predicted which is basic requirement under the LGR procedure.

What would be the policy level stance for treating homophones? Will they be treated the same as "color" and "colour"? I mean "हिंदी" and "हिन्दी" would potentially be registered as separate domain names, right?

Regarding Halant ending words:

This we can accommodate as the ending halant in many cases is not clearly visible. Just like Santhali variant cases, these can be missed by users by not expecting them to be. Request all for a feedback on the same.

As far as the Nepali language is concerned, it does not matter much in domain names whether to include a "halanta" or not a "halanta" at the end of a word since they would not distort the intended meaning of the word, for example, "सांसद" vs "सांसद्". So our recommendation would be to treat them as "blocked variants", i.e., it should be okay to accept either of them as a domain name and restrict the other. Regards, Bal Krishna

Hello Akshat and All, Please find attached the inputs from the Devanagari Nepali and Newar Languages for the Devanagari LGR Report. Regards, Bal Krishna

-- Regards, Akshat Joshi C-DAC GIST

------------------------------------------------------------ ------------------------------------------------------------------- [ C-DAC is on Social-Media too. Kindly follow us at: Facebook: https://www.facebook.com/CDACINDIA & Twitter: @cdacindia ]

This e-mail is for the sole use of the intended recipient(s) and may contain confidential and privileged information. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies and the original message. Any unauthorized review, use, disclosure, dissemination, forwarding, printing or copying of this email is strictly prohibited and appropriate legal action will be taken. ------------------------------------------------------------ -------------------------------------------------------------------