- Latingp - lists.icann.org

Re: [Latingp] Homoglyphs within Latin script
by Tan Tanaka, Dennis Jan. 4, 2018

Jan. 4, 2018

Mats, Are you saying that the panel needs to do more analysis to determine which one to include in the repertoire or is the matter that both of them will be included but you are suggesting some variant rules? On the latter, IP briefly discussed about the two code points in the feedback regarding the Principles document. -Dennis From: Latingp <latingp-bounces(a)icann.org> on behalf of Mats Dufberg <mats.dufberg(a)iis.se> Date: Sunday, December 31, 2017 at 12:41 PM To: "latingp(a)icann.org" <latingp(a)icann.org> Subject: [EXTERNAL] [Latingp] Homoglyphs within Latin script There are many code points that are so similar that confusion is at risk, but there is at least one clear pair of homoglyphs: 0259 ə LATIN SMALL LETTER SCHWA 01DD ǝ LATIN SMALL LETTER TURNED E They are not homoglyphs in upper case. In two of the languages that attest either of the two, it is not possible to say which because there is no upper case letter. Mats --- Mats Dufberg DNS Specialist, IIS Mobile: +46 73 065 3899 https://www.iis.se/en/

2 1

Homoglyphs within Latin script
by Mats Dufberg Dec. 31, 2017

Dec. 31, 2017

There are many code points that are so similar that confusion is at risk, but there is at least one clear pair of homoglyphs: 0259 ə LATIN SMALL LETTER SCHWA 01DD ǝ LATIN SMALL LETTER TURNED E They are not homoglyphs in upper case. In two of the languages that attest either of the two, it is not possible to say which because there is no upper case letter. Mats --- Mats Dufberg DNS Specialist, IIS Mobile: +46 73 065 3899 https://www.iis.se/en/

1 0

Latin character repertoire
by Mats Dufberg Dec. 29, 2017

Dec. 29, 2017

Friends, This is a status report of my processing of the repertoire document(*). It takes longer time for me to process than I thought, but I still hope to have it done before next meeting. As I write this, row 336 is the last line I have processed. That line is interesting because the code point is attested, but it is *not* in MSR2. It was incorrectly added and marked as in MSR2 and had wrong description. This is a *new* case, not any of the cases we have looked at before. 1E3D ḽ LATIN SMALL LETTER L WITH CIRCUMFLEX BELOW I have, as stated above, processed up and including row 336. In column B you will see the "final review", which is the result of my review. If it says "ok" everything looks well and the code point (sequence) is OK to include. If it says "no attest" none of the languages that we have investigated use that code point (as a character). I have then copied the row to tab "removed from main data" and stricken through. These case should not be a problem. If it says something else in that column, it could be worth looking at. In some of those cases I have consider it to be not attested and done as with "no attest". If anyone can look at all rows before row 336 which has neither "ok" nor "not attest" in column B it would be a help to improve quality. Yours, Mats *) https://docs.google.com/spreadsheets/d/1Sj2moWEw4YFLFTcvsiRpjsQRyLZJ7SCnlF7… --- Mats Dufberg DNS Specialist, IIS Mobile: +46 73 065 3899 https://www.iis.se/en/

1 1

Re: [Latingp] Latin-Armenian cross-script variant analysis
by Tan Tanaka, Dennis Dec. 27, 2017

Dec. 27, 2017

Dear Igor, Thank you for your detailed response. I believe we have what we need to start our analysis work on Latin and Armenian. We very much appreciate your assistance. Best, Dennis From: Igor Mkrtumyan <imkrtumyan(a)isoc.am> Reply-To: "imkrtumyan(a)isoc.am" <imkrtumyan(a)isoc.am> Date: Wednesday, December 27, 2017 at 8:23 AM To: Dennis Tan Tanaka <dtantanaka(a)verisign.com>, "ArmenianGP(a)icann.org" <ArmenianGP(a)icann.org> Cc: "'Tan Tanaka, Dennis via Latingp'" <latingp(a)icann.org>, 'Pitinan Kooarmornpatana' <pitinan.koo(a)icann.org>, Sarmad Hussain <sarmad.hussain(a)icann.org> Subject: [EXTERNAL] RE: Latin-Armenian cross-script variant analysis Dear Tan, The most common Armenian fonts used in Internet applications are Arian AMU, Arial, Times New Roman, Sylfaen. Their Armenian small H did look like a Latin Small H. I am attaching two screenshots from the most popular Armenian Internet media sites to convince you that they are homoglyphs. They set de-facto standard for the most of the Internet applications. At the same time I admit that there can be other fonts but they are rarely used. Thank you Igor From: Tan Tanaka, Dennis [mailto:dtantanaka@verisign.com] Sent: Monday, December 18, 2017 7:49 PM To: imkrtumyan(a)isoc.am; ArmenianGP(a)icann.org Cc: 'Tan Tanaka, Dennis via Latingp'; 'Pitinan Kooarmornpatana'; 'Sarmad Hussain' Subject: Re: Latin-Armenian cross-script variant analysis Dear Igor, As a follow-up to my previous email. One of the Latin panel members noted that he sees a different type face using the same compart.com URL. He was using a Win10 computer, whereas I was using a Mac. In his case, U+0570 did look like a Latin Small H, and the type face in-use was DejaVu Serif (http://dejavu.sourceforge.net/samples/DejaVuSerif.pdf, see page 11). I believe this is in-line with what the Armenian panel was dealing with. In light of this information, let me tweak my earlier question: - What type face(s) (i.e. fonts) is/are in current and regular use by internet applications for the Armenian script? Thank you, Dennis From: Dennis Tan Tanaka <dtantanaka(a)verisign.com> Date: Friday, December 15, 2017 at 9:37 AM To: "imkrtumyan(a)isoc.am" <imkrtumyan(a)isoc.am>, "ArmenianGP(a)icann.org" <ArmenianGP(a)icann.org> Cc: "'Tan Tanaka, Dennis via Latingp'" <latingp(a)icann.org>, 'Pitinan Kooarmornpatana' <pitinan.koo(a)icann.org>, Sarmad Hussain <sarmad.hussain(a)icann.org> Subject: Re: [EXTERNAL] RE: Latin-Armenian cross-script variant analysis Dear Igor, Thank you for your prompt response. I hope this email gets delivered with the images as it is better that way to convey what we are trying to achieve. Let’s use Armenian Small Letter HO (U+570) as an example. Below, you should see the three visual representations from 1) compart.com<https://www.compart.com/en/unicode/U+0570> (the URL you recommended), 2) Unicode<https://unicode.org/cldr/utility/character.jsp?a=570&B1=Show> and 3) The Armenian LGR proposal. [cid:image001.png@01D37EF2.35E98C40] [cid:image002.png@01D37EF2.35E98C40] [cid:image003.png@01D37EF2.35E98C40] Observations: - The visual representations (i.e. type face) from compart.com and Unicode match, but the representation in the Armenian LGR proposal is different. - The Armenian LGR proposes Latin Small Letter H as variant of Armenian Small Letter HO because they were found to be homoglyphs per the panel. - The Latin GP wishes to learn what are the common type faces used in internet applications, such as browsers in order to do our variant analysis. - Question: is the following premise true? The type face used by compart.com or Unicode (shown above) is not representative of current use by internet applications. - Question: which type face(s) does the Armenian GP used in your LGR document as seen above? And, is the type face used in Armenian LGR proposal representative of current use by internet applications? I hope this is helpful. Alternatively, I can jump into a quick call to explain all this. Thank you. Best, Dennis From: Igor Mkrtumyan <imkrtumyan(a)isoc.am> Reply-To: "imkrtumyan(a)isoc.am" <imkrtumyan(a)isoc.am> Date: Thursday, December 14, 2017 at 6:53 AM To: Dennis Tan Tanaka <dtantanaka(a)verisign.com>, "ArmenianGP(a)icann.org" <ArmenianGP(a)icann.org> Cc: "'Tan Tanaka, Dennis via Latingp'" <latingp(a)icann.org>, 'Pitinan Kooarmornpatana' <pitinan.koo(a)icann.org>, Sarmad Hussain <sarmad.hussain(a)icann.org> Subject: [EXTERNAL] RE: Latin-Armenian cross-script variant analysis Dear Dennis, We are very glad to receive a letter from you and the Latin GP. It signifies that we can finalize soon our Proposal for an Armenian Script Root Zone LGR. The most commonly used Armenian type face (based on current use on Internet applications, such as websites and browsers) can be seen here https://www.compart.com/en/unicode/scripts/Armn . It differs slightly from the Unicode default. However we are ready to discuss your objections and come to an agreement. What code point seems doubtful to you to be considered as a homoglyph? Thank you and best regards Igor Mkrtumyan Armenian GP From: Tan Tanaka, Dennis [mailto:dtantanaka@verisign.com] Sent: Tuesday, December 12, 2017 2:23 AM To: imkrtumyan(a)isoc.am; ArmenianGP(a)icann.org Cc: Tan Tanaka, Dennis via Latingp; Pitinan Kooarmornpatana; Sarmad Hussain Subject: Latin-Armenian cross-script variant analysis Dear Armenian GP, The Latin GP is initiating work on Latin/Armenian cross-script variant analysis. To this end, we want to use the Proposal for an Armenian Script Root Zone LGR, version 7 of 2015-11-05 as a reference. In your proposal, there are seven cross-script Latin/Armenian variant sets. However, we cannot replicate your results for the exception of one code point. The reason for this is the choice of type face. Our wg cannot find the type face the Armenian GP used to do the visual comparison. We kindly ask your help in understanding the following: · What is the type face (i.e. font) that the Armenian GP used in their analysis? · Was the decision of using a different type face different that the Unicode’s default (i.e. type face used in Unicode code charts) based on current use on Internet applications, such as websites and browsers? Please elaborate to the extent possible. Thank you in advance for taking the time to answer this email. Sincerely, Latin GP, variant group

1 0

Happy holidays!!
by Mirjana Tasić Dec. 23, 2017

Dec. 23, 2017

Dear Latin GP members I wish you a very happy Holiday season and a wonderful New Year Regards Mirjana

1 0

Latin-Armenian cross-script variant analysis
by Tan Tanaka, Dennis Dec. 18, 2017

Dec. 18, 2017

Dear Armenian GP, The Latin GP is initiating work on Latin/Armenian cross-script variant analysis. To this end, we want to use the Proposal for an Armenian Script Root Zone LGR, version 7 of 2015-11-05 as a reference. In your proposal, there are seven cross-script Latin/Armenian variant sets. However, we cannot replicate your results for the exception of one code point. The reason for this is the choice of type face. Our wg cannot find the type face the Armenian GP used to do the visual comparison. We kindly ask your help in understanding the following: * What is the type face (i.e. font) that the Armenian GP used in their analysis? * Was the decision of using a different type face different that the Unicode’s default (i.e. type face used in Unicode code charts) based on current use on Internet applications, such as websites and browsers? Please elaborate to the extent possible. Thank you in advance for taking the time to answer this email. Sincerely, Latin GP, variant group

2 3

Canceled: Repertoire WG meeting - alternate Tuesdays - 5pm UTC
by Pitinan Kooarmornpatana Dec. 18, 2017

Dec. 18, 2017

How to connect: Link to Adobe Connect: https://participate.icann.org/idn/ *Upon logging into Adobe Connect, a pop up window will provide you the option to Dial Out to your Phone. Enter your Phone Number* (Remember to change the Country Code if needed). After joining the call, as a courtesy to others and the presenters, please MUTE your phone. This can be done by selecting *6 on your keypad. To UNMUTE select *6 again. If you are Unable to log into Adobe Connect and can only join via phone: https://www.myrcplus.com/cnums.asp?bwebid=8369444&ppc=1064158692&num=1-719-… Participant Passcode: 1064158692

1 0

Latin GP proposal for inclusion letters in MSR
by Mirjana Tasić Dec. 14, 2017

Dec. 14, 2017

Dear IP members, Please find enclosed the Latin GP proposition for inclusion some Latin script letters in new version of MSR. Best regards Mirjana Tasic Latin GP chair

3 2

Tongan evidence for line 62 of Latin GP Code Points
by Meikal Mumin Dec. 14, 2017

Dec. 14, 2017

Dear Mats, in the shared Dropbox I created a folder Tongan, which contains a file called "tongan-lg.pdf". On pages 14-15 you can find further details about the orthography. Long vowels are represented in Tongan orthograpy by use of a Macron ("long vowels, as in màlò"). Meanwhile definiteness is marked by stress, which is is written in Tongan using a rising accent ´ following the vowel ("The fakamamafa pau or defi nitive accent falls on the fi nal vowel of a word and designates defi niteness, as in tangata ́ (the man). An example of the difference that the defi nitive accent can make is this: ko e tangata means “a man” but ko e tangata ́ means “the man”.") Logically this leaves us with a sequence ā´, which should be confirmed by an expert on the language. Meanwhile Wikipedia <https://en.wikipedia.org/wiki/Tongan_language#Tongan_alphabet> informs us that "Although the acute accent has been available on most personal computers from their early days onwards, when Tongan newspapers started to use computers around 1990 to produce their papers, they were unable to find, or failed to enter, the proper keystrokes, and it grew into a habit to put the accent after the vowel instead of on it: not á but a´. But as this distance seemed to be too big, a demand arose for Tongan fonts where the acute accent was shifted to the right, a position halfway in between the two extremes above. Most papers still follow this practice.". This is why I commented this in column J "Possible variant of ā´ ". I hope this clarifies the situation. Best, Meikal

4 4

Proposal for inclusion some Latin script letters in MSR 3
by Mirjana Tasić Dec. 12, 2017

Dec. 12, 2017

Dear GP members Please find the link to the document with proposal for inclusion some letters in new version of MSR. https://drive.google.com/open?id=1Hk0Y--AZZrbL6a7cxAHEWywjU9k6Jh72rj5gcZtRb… Please, comment and correct ! Regards Mirjana

1 0