Closed neogitter closed 2 years ago
I do understand that if the baseline of Telugu is lowered anymore than the current level the conjuncts that have descending strokes will move further down and will be chopped off at the bottom. So is reducing the overall size of the font a possible solution?
Thanks for your bug report! Date/time formatting in Android, and the use of the ‘&’ character, is unrelated to fonts; I’ve forwarded your report to our localization team [tracking number: b/29672833].
@roozbehp, what do you think about the Telugu baseline in Android? According to this README.android file, Android is scaling the Malayalam and Tamil UI fonts to 90%; should this also be done for Telugu?
Actually, do you know if there’s a particular reason why the Noto UI fonts aren’t already released in a form that works for Android without further modification?
@brawer There are lot more bugs in Telugu language at OS level. May I please know the proper channel to report them?
The words are actually నాకు గుర్తు లేదు , సూర్య వంశస్తులు , స్వార్థపరుడు . See how the lower parts were cut off in https://translate.google.com/community on Android mobile. Due to this it's difficult to understand the exact words for things like సై స్తై స్తె etc. Hope you will find a solution.
@neogitter, thanks for your reports! Regarding the clipping on translate.google.com, I’ve forwarded your bug report to the Google Translate team. By the way, the Chrome address/search bar has a similar clipping problem; see Chromium bug 623492. Regarding Telugu localization in general, I’ve asked the Google localization team how you could best reach them, but haven’t heard back yet.
@neogitter, just heard back from Google’s localization team. For Android bugs with Telugu, please continue to file bugs similar to what you’ve done with bug 214230, it’s the perfect way to reach them. So let’s restrict this bug here to the font problems.
@brawer Thanks. Let's see if the clipping problem for Telugu can be fixed in time for N release hopefully. So did they decide to clarify down Telugu to 90% too?
Here the problem is much more obvious. Google indic keyboard. All the conjuncts are clipped.
Actually, do you know if there’s a particular reason why the Noto UI fonts aren’t already released > in a form that works for Android without further modification?
In case of Tamil, even without scaling, there's no clipping. Scaling was done to make the line spacing more comfortable IIRC.
As for Telugu, the UI version should fit in without clipping. If there's clipping even when Roboto's default line-spacing is used, we should look at the metrics and investigate why.
Many different issues here. They should be tracked separately. For example, the very last screenshot is about horizontal clipping, not vertical clipping. I don't know if it's the keyboard app or the font app either.
As for vertical clipping in Google Translate, that's a bug that needs to be filed against Google Translate. They seem to be doing their own clipping.
As for the baseline shift, it seems unavoidable. We tried a lot of back and forth with the font designer. He tried a lot, and this is the best he could come up with.
As for scaling the font 90%, I don't see why unless there's a similar concern to Malayalam and Tamil, where although there was no clipping, the vertical space between consecutive lines was too tight.
@roozbehp The baseline shift becomes really ugly when using Telugu along with numbers. The difference in baseline becomes glaringly visible and is quite awkward. Wouldn't it help if we scale down the font to 90% and shift the baseline down to match English?
The horizontal clipping is seen in the official Google infic keyboard app and also in some messenger apps, eso the last characters.
Android N & the new version of Google Indic IME solved a lot of problems. However, the baseline shift between Telugu & Alphanumeric char still happens.
While I was verifying this issue, I noticed OSX actually uses the same baseline with their Telugu font & the alphanumeric chars. Can we re-reconsider to use the same baseline? Please refer to the attached screenshot.
My guess is this is NotoSansTeluguUI, and in order to accommodate descenders within the UI constraints, the baseline for the Telugu ends up being higher than that used by Roboto, which I guess is what is being used for the digits and Latin. Either the descenders get crushed, or Latin gets added to NotoSansTelugu (maybe all UI variants of Noto) and Android figures out how to use it, or we live with this.
Is it me or does OSX scale the Latin down so that it's on the same baseline? In this case perhaps a special latin could be added to Telugu UI
When I type the same thing inside a website and in the address bar in Chrome the outputs are different. The base line is correct in the website but not address bar.
Re '&' in Android in Telugu, I can't see it in any of the screenshots. Where are you seeing it?
The date and time formatting problem was from CLDR. The "[AM]" problem was already fixed. But there is still a bug in day of week order. I just filed that as http://unicode.org/cldr/trac/ticket/9952.
The only remaining bug in this issue seems to be the baseline issue, at #1
I would like to bring to your notice about the bugs in Telugu language at OS level.