Monday, May 05, 2008

Reply to Anivar

It seems like there is some problem with comment posting on Anivar's blog. So I decide to put my reply as a separate blog post. Here it goes:

Few factual corrections and comments:
1. About the pango bug 357790 and the patch on it:
The patch on this bug is a mere clean up version of the patch on bug 121672 which was originally created by LingNing. This was also based on the inputs given by Ani about the grammar of 0d30 and 0d31 which was later resolved (to 0d30 only) through discussions with smc.
Point is not to transfer the responsibility, but to acknowledge that pango genuinely has a problem that it does not behave the way Uniscribe does. Another problem in this case is that, Uniscribe bahavior has changed from its version in XP to Vista and we are yet to fix this bug completely. Anyway, my patch was reverted one year back (see Comment #32 on bug 357790). Ever since then I have urged on concentrating on the original issue which I still continue to. And Lohit was agreed to be fixed temporarily to suite the current pango.

2. About the size details of Meera font:
First, Pravin compared Lohit vs Meera (not Rachana vs Meera). But I hope this was only a minor typo.
Second, as can be seen from the comments on Pravin's blog, he is not the only one to complain about this issue.
Third, the Ascent-Descent and Em-size you mentioned here defer from the font that was given for inclusion in fedora. The font given for fedora has Ascent=1147, Descent=901 and Em Size=2048.

3. About Hiran's fixes and bug reports:
Reporting a specific bug is always welcome , but comments like 'a font with 1000 bugs' and 'unfixable font' are the ones that I "personally" felt were 'agitation', although it is not necessarily a negative term.
Thanks to Hiran for his contributions towards lohit and samyak, and I still encourage him to do that. But there are reasons for not accepting his fixes directly. I think they should be understood in detailed and learned as a student to be more efficient in an open source development environment. The reasons are:

[i] Hiran's font has problem in its header section due to which every time the font is open for editing in fontforge, a notice is thrown out that the font has 'restricted license' and you need legal permissions to edit it.
[ii] The font provided by Hiran as a fixed version did not appear to be based on the latest version of .sfd in the cvs (now svn) repository.
[iii] I was not sure of all of his intended changes. As in any other project, Lohit needs to put a Changelog for every commit. So I had to copy the changes based on his email(I regret not forcing the bugzilla then) and the visible fixes in the ttf file.
[iv] Not being based on the original .sfd and not being submitted as an .sfd, there was possibility of cvs conflicts to occur. So it was necessary to create a cleaned diff patch out of his fixed font so that changes could be both recorded and committed without any hassles in the repository.
[v] I should have made him aware of all these factors, but I thought it would be kind of discouraging for him to do all the procedural details, so I did it on my own. May be I was wrong on this. I might post a detailed description of 'How to contribute to open source font projects' sometime soon.

4. About the authority:
I have not received any authority by just claiming it. If you are active on Indlinux community and other upstream mailing lists like gtk-i18n, gnome-i18n etc. you would know that I have got it (if any) through contributions only. My contributions are not limited to only one language but they span through entire Indic spectrum. Not that I over-estimate myself or under-estimate anyone, but do you have anything 'real' against me apart from bug 357790 to say that I have done harm to Malayalam? If one arguably incorrect fix means harm to language, what is your opinion about mistakenly done errors by other members. See bug 504810. It is evident here that the patch for Malayalam is mixed with section of Sinhala, probably a typo. Fortunately there is no side-effect of this detected so far. I wont even blame Praveen A, for mistakenly removing ZWJ flag that he corrected here, I wont say he has "harmed" Malayalam for about an year by doing so in his earlier patch. Things happen and you don't have to see things as black and white. Anyway if you are aware, we already had huge amount of discussion over '0d30' issue too. I may not even raise question about SMC's authority though and leave it up to the people of Kerala.

5. About the smc-fonts inclusion:
I don't know whether to call it duty or anything else. But those who attended my talk in foss.in last year know that including more fonts in fedora from community and for community was on my list of plans this year. We have managed to do that for few other languages as well like sarai-fonts for Hindi, madan-fonts for Nepali and few Thai fonts as well. Idea has been to give more options to the users and there is still a lot do on this plan.

6. Thanks for notifying the incorrect link on www.fedorahosted.org/lohit. It is corrected now.

7. The bug fixing on Lohit started long back in 2006. But continuous changes in the requirements given by various people (bugzillas have all the records), changes in pango, changes to OpenType and recently the changes required in font with reference to my 1st point above, the bugs were ought to be popped up every now and then. Thanks again for notifying the current rendering issues, which are mostly recent. I would appreciate to see them in bugzilla too. Anyone who wishes to do that please refer to i18n bug guidelines. Reading them wont be much effort as they are not too many, but they are certainly worth following.

8. Thanks for documenting the design issues in Lohit. In addition, I would also mention that the kerning and spacing between glyphs is not accurate, which is a major issue IMO that the font looks bad when viewed in paragraphs. This probably was introduced by the original designers when they resized the font to match the readable size.

Hope, we work with more cooperation and harmony.