Foreign Word Detection, Malayalam Writing Learning Portal and More: SMC Monthly Updates October 2020
Foreign Word Detection in mlmorphPython library for Malayalam morphological analyzer - mlmorph released version 1.
SMC has to clarify its position in regards to the concerns raised on RITs statement posted in Rajeesh KV's blog and other instances of same statement here and here. The statement came as a surprise to the project admins of SMC since this statement was made by key members of SMC project, specifically Rajeesh KV and Hussain KH. The claims run deep into several key areas of how SMC operates as a Free and Open Source Software organisation and the notions of copyright. As any organisation that operates out in the open, SMC project's secretary, Hrishikesh, initiated a discussion in the public mailing list, and members volunteered to create a task force that would investigate these claims and demands, and then recommend a way forward.
Volunteers for the task force spanned old and new members. They are :
Following is a report of the team examination.
The development of various Unicode fonts for Malayalam language is one of the most fruitful and satisfying achievements of SMC's community members. Our focus is on creating free and open source technology building blocks for Indian languages. We are not just a typical font foundry that develops and releases fonts. Our font development process also builds technology to enable font development, with the notion that others (including private companies) will find them as useful references to build new fonts, across languages. This encourages the growth of a strong, standards-compliant, uniform language ecosystem for free and open fonts and that is beneficial for everyone. For this reason, all our fonts are released under Free Software licenses (OFL /GPLv3 with Font Exception). We expect people will re-use the technology to build more free and open fonts, adhering to corresponding free Licenses. When there are innovations and improvements made while developing new fonts, these often get back-ported to our older fonts as well. We are thankful to all the contributors who have made this possible. More on history of fonts here. We love to see evolution of new Institutions taking up the need of building more fonts. We understand many Foundries may have issues in understanding Free and Open font Licensing, we request you to write to us for any clarifications at email@example.com
This refers to point 6 stated from RITs statement in Rajeesh KV's blog , reproducing the relevant parts below.
"In 2015, Santhosh Thottingal split the comprehensive lookup rules from the Fontforge SFD file of Rachana into a separate feature file, but the copyright statements were not preserved . It is our opinion that removing copyright statements is violation of copyright act (hence a crime) and immoral in the free software world. This is also the root cause for missing copyright in the OpenType lookup rules and build script, of the fonts in question."
Each font maintained by SMC is a separate software package derived from a dedicated code repository that includes various font components - including design files, look up tables, tests and build tools. The copyright and license is also maintained in a LICENSE and/or AUTHORS file in same repository. The said feature file separation was done within the repository, hence there was no obvious need to separately mention either licensing or copyright in individual files. This is standard practice adopted in most free and open fonts. One needs to specifically mention the licensing only if Licensing/Copyright Statements are different from the license/copyright of the overall repository.
In short, Rachana's feature file inherits the licensing and copyright information from its License file within same repo and this should be clear for any free software developer. It is not a violation of copyright act/crime/immoral , and can be classified as a misunderstanding of licensing from Rajeesh KV, who reviewed the same code in that commit.
In addition to all this, the changes (splitting the lookup rules) were carried out in the open, not only code wise, but also via emails including both Hussain KH and Rajeesh KV, and were widely applauded at the time by members of the community. In that context, criticising the main author of the change in harsh words is unfortunate and unjust.
This refers to point 12 in RITs statement in Rajeesh KV's blog, reproducing the claim here.
"The Fontforge based ‘build script’ added by Santhosh Thottingal used to generate TTF/OTF/WOFF/WOFF2 files is adapted from that of Amiri font by Khaled Hosny  without preserving copyright or attribution. RIT requests to credit the original author[s] of this tool. It is of our opinion that removing copyright statements from a free software program code is illegal and immoral. It is also hypocritical when a person who asserts one’s own credit does this crime to other well-known and respected free software developer[s]."
See gitlab issue#9 for discussions on this claim.Original claim and subsequent discussion on GitLab reveals the following
"The person who associated a work with this deed has dedicated the work to the public domain by waiving all of his or her rights to the work worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law. You can copy, modify, distribute and perform the work, even for commercial purposes, all without asking permission."
We conclude that this claim is invalid, and was made without ample research. Considering that the authors are knowledgeable in free software licensing and copyrights, this seems to be an attempt at mudslinging.
Further, point no. 14 of the statement amplifies this character assassination attempt by adding on allegations of plagiarism, when Santhosh reproduced some content he added in SMC wiki in a personal blog-post.
As per point 13 in RITs statement in Rajeesh KV's blog,
"Test cases in the ‘test file’ are contributed by various contributors, RIT request to add the attribution of such contributors to the extent possible (Kavya Manohar, Santhosh Thottingal, Rajeesh KV). Santhosh Thottingal has responded to this request as “test cases were mainly prepared by Kavya Manohar and no need to have attribution”, but RIT firmly believes the copyright statements of the contributors must be added."
See Gitlab issue#13 for discussions on this claim. The discussions reveal the following.
1. The current test data is derived from an initial version prepared by Kavya Manohar and improvement of same through various bug reports.2. Usually these are given credit in a file - copyright is given for substantial contributions, but bug reports are usually counted as minor contributions.3. Since test cases are handled in repository level minor contributions, separate License or copyright assignment is not needed unless a separate license is specified by in the repository.
If it is separated out and reused, SMC concludes that the contributions to the tests should be acknowledged in contributors list, unless the contributors make a contribution under a separate license. In such cases, the authors of the repository are obligated to review and include the files under those licenses and mention it in contributors list/a separate file as they see fit.
Point No 7 hints that there is missing attribution to authors of the font and point No 11 says following
"Preserve the copyright of the original authors of the ‘lookup rules’ and Naming convention (notation for Glyphs) in all these fonts."
See issue#12 for discussions on this claim.
Keeping the history of fonts and font maintenance above we conclude the following on this issue.
1. All fonts that directly used look up rules developed for Rachana or used rules that were derived from it, were worked on by the same set of maintainers. Suresh P, Rajeesh KV, Santhosh Thottingal and Kavya Manohar were the major contributors to the maintenance effort, with others joining in later on the projects' history.
2. The opentype rules, which were derived from initial commits on Rachana, had several refactors on other fonts, and many of them were ported back to Rachana and other fonts as well over the time. When the rules were first ported across fonts in 6.0 release , nobody ever thought about the need of a procedural attribution based on first commit code license, throughout fonts which are part of same release (under various licenses). This was because the maintainers of Rachana are active maintainers of SMC fonts.
3. We firmly believe there is no copyright for Naming Convention (notation for glyph naming) as claimed on Point No 11 in RITs statement in Rajeesh KV's blog. We consider naming conventions as a practice to follow and it should be based on a documentation of the practice. We value and recognise immense contributions of Hussain KH on building and standardising Malayalam free and open fonts ecosystem and restoring traditional script in usage. We think this copyright claim on naming conversion is misplaced and would like to address it with increasing licensing awareness within community and implementation of a proper attribution policy.
To explain further, all fonts we maintain are copyrighted to their respective designers and engineers. If they use Rachana's lookup rules (modified or otherwise) the font sources must retain a file that mentions that the font contains work that falls under the following copyright.
"Copyright: Digitized data copyright (c) 2004–2005 Rachana Akshara Vedi (Chitrajakumar R, Hussain KH, Gangadharan N, Vijayakumaran Nair, Subash Kuraiakose), (c) 2006–2016 Hussain KH, Suresh P, Santhosh Thottingal, Rajeesh KV, Swathanthra Malayalam Computing (http://smc.org.in)".
The creators are free to select OFL1.1 or GPLv3+FE as the license of this component. The resulting font should be licensed under a compatible License to the component licensing. This does not mean that the fonts themselves are copyrighted to the creators of Rachana, but that the fonts contains a part of work copyrighted and released under free licenses by authors of Rachana.
We understand where RIT's confusions are originating from. It is about enabling authorship. We understand our licensing enforcement request in 2014 to ATPS within community discussion list which seems to have created confusion among few.
We specifically requested following at that time
1. Comply with the original licensing of the base font
2. Preserve copyright attribution, ie, mention the license and attribute the authors of the original parts that were reused.
It was not a copyright claim on derivative components, but more a request for derivative font to comply with the licensing of the original. As an endorser of free software ecosystem, SMC believes in enabling creativity, authorship and collaboration on Open font building. We think a font's copyright should be held by the designer of glyphs and the open type engineers who maintain it. But the font licensing should be based on the license of the building blocks like lookup tables, build scripts etc and each component used should be attributed clearly. SMC should soon add a best practice documentation for this purpose.
This incident and the resulting discussions lead us to recommend the following steps.
We noticed that the statement made by Rajeesh KV and Hussain KH et al. = who are people who have played critical role in the development and growth of font infrastructure - contains claims that are phrased as personal attacks. SMC has a public mailing list where grievances can be aired. We do have a code of conduct , and it specifies a contact e-mail where any concerns can be raised. This should be the ideal mechanism to resolve conflict between members if it cannot be resolved by friendly means. We make it clear that the code of conduct mandates that personal attacks should be avoided and any grievances that SMC members has with each other should ideally be resolved through discussions, and involve the project admins.
We reiterate that the statement includes unsubstantiated personal attacks against voluntary contributions of Santhosh Thottingal who has been a prolific contributor member of SMC. We request Rajeesh KV and Hussain KH to amend their published statement in light of the facts above and to stick to reasoned criticism and constructive comments instead of personal attacks henceforth. We hope RIT understands that the "Conclusion" section in their statement is misplaced and request RIT to do needed amendments to the statement, especially by removing the following misleading section in light of the clarifications.
" This surrounds the fonts by SMC and RIT in Fear, Uncertainty and Doubt (which the corporate proprietary companies successfully used against free software for years) and put all the individual users and organizations and developers using these fonts under legal risk and moral ambiguity."
We invite RIT to be an institutional contributor in SMC's font project community, and use project's public channels for discussions and disputes rather than depending upon personal emails, individual judgements and one sided public statements.
In the course of following up on the statement from Rajeesh KV and Hussain KH, it became clear that the underlying cause is not copyright/credit/attribution dispute. The statement appears to have been born out of spite from unresolved conflicts between members over misunderstandings. Further introspection reveals that the motive behind making claims in the statement would not have been born had we been kinder to each other and be more open in communication. The number of face to face meet-ups between SMC members have gone down significantly, and that affects the organization's momentum because of a lack of real world, human, connections. We recommend that SMC as a community have regular meetings (quarterly at least) organized. We also recommend that SMC as a group of humans who are friends to come together and help resolve these issues so that such incidents are avoided. We hope anyone who can't resolve a question or conflict will reach out to us directly via firstname.lastname@example.org or via project's public channels so the larger community can be involved and resolution can be achieved without mudslinging.