Grants talk:Project/slevinski/ASL Wikipedia 2-D Font Development for SignWriting/Archive 1

From Meta, a Wikimedia project coordination wiki
Jump to navigation Jump to search

Comments[edit]

  • Hi Slevinski and Valerie. Thanks for your proposal on font development work for ASL Wikipedia on Incubator. I've talked about this proposal a little further with some of the grant officers on our team. I think you make a fair case that, in order to build a Wikimedia project that can be written in sign language, a font needs to be developed and rendered just like it is for other fonts that appear on our projects for other languages. However, one concern I have heard and agree with is that it is unclear who is likely to use this font for the purpose of improving ASL Wikipedia. Right now, the community engagement plan reads:
There are a variety of websites and email lists that can be contacted before the project starts and after the project is completed. Additionally, once the font is available and used on the various projects, it will organically grow as users are exposed to the font and the new reality.
It's important that the grants that are funded have outcomes that are likely to improve Wikimedia projects in some manner. I think as a first step, it'll be important to demonstrate, in this proposal, that there is a audience interested in editing or improving ASL Wikipedia (or creating another Wikimedia project based on some other form of sign language) should this font become available. Providing a list of some websites as you mentioned will also be useful, but I think that the Project Grants committee will want to see some evidence that there is interest from ASL speakers in this project. If you've done outreach with any ASL communities over e-mail or otherwise, and they've already provided feedback on this project, you can simply paste that feedback on the talk page here when you get a chance. Thanks, I JethroBT (WMF) (talk) 01:40, 14 March 2017 (UTC)
  • Hi JethroBT and Slevinski. Thank you, JethroBT, for your response and attention to this important project for writing sign language Wikipedias. And thank you, Slevinski, for developing the software that makes it possible for our current "test" Sign Language Wikipedias posted in the Incubator right now: American Sign Language Wikipedia: https://incubator.wikimedia.org/wiki/Wp/ase and the Tunisian Sign Language Wikipedia: https://incubator.wikimedia.org/wiki/Wp/tse.
Your comments JethroBT, bring up the issue of the "cart before the horse". Let us imagine that the "cart" is the easy-to-use editing interface needed to encourage sign language users to write their articles in their sign languages. The "horse" is the Unicode font and resulting improved software development needed, that will make it possible to write sign language articles on phones, tablets and computers in an easy way.
Steve Slevinski and Yair Rand (java-script SignWriting keyboard) developed a "test cart" to show the world that it is "possible" to write articles in written sign language. "Written Sign Language", or the "SignWriting Script" is one of the world's established scripts, but it is new to signers - The ASL Wikipedia is an amazing experience for signers because they never knew their language could be written before, and when they start reading the articles posted in the ASL Wikipeida, they are learning for the first time how to read their written sign language.
At this moment, Deaf authors who know our software, Adam Frost, Holly Sharer and Jason Nesmith, have been contributing articles to the ASL Wikipedia on the Incubator...they know SignWriting and the current software well, but they also told me that their Deaf friends want to write articles but their Deaf friends find the software as it stands right now, too difficult and they are waiting for the new software for phones and tablets (which requires the Unicode font described in this grant request).
Meanwhile, even with the "cart before the horse" position we are in, we still have around 60 articles in the ASL Wikipedia on Incubator. Just yesterday Holly Sharer wrote a new article in ASL about CATS, and she also edited and added to another article she wrote last month... So there is activity in the ASL Wikipedia, but everyone is waiting for the Unicode font.
Here is the article on CATS posted yesterday with some editing today:
https://incubator.wikimedia.org/wiki/Wp/ase/M569x521S2ff00482x483S20e00540x501S26a06555x494S1d410511x488
Here is a booklet showing excerpts of some of the ASL Wikipedia articles in a PDF:
http://www.signwriting.org/archive/docs12/sw1196_ASL_WIKIPEDIA_October_2016.pdf
As soon as we have the Sign Language Wikipedia user-interface for phones and tablets using the Unicode-based SignWriting font, I plan to contact schools for the Deaf around the United States and English-speaking Canada, where ASL is in the schools, to suggest to teachers to make it a Deaf classroom project to start writing articles in the ASL Wikipedia on their tablets and phones.
So you have our California non-profit organization behind the dissemination and spreading of the writing of Wikipedia articles, but before the right tools are available we are stuck, and our non-profit, like Wikipedia in general, is volunteer, except for software development, which we need funding for...I am a volunteer, but I promise that our organization will be working for the spreading of writing articles in Sign Language Wikipedias as soon as we have the right font. We have been working on this project for many years and we have never given up.
The truth is, without the ASL and other Sign Language Wikipedias, it is hard for signers to grasp what it means to write their languages. Sign languages have never been true written languages before...and the Sign Language Wikipedia project is an outstanding way to teach Deaf and hearing signers how to read their "written" language. --ValSutton (talk) 16:47, 14 March 2017 (UTC)

Expand Grant to Teach Classrooms of Brazilian Deaf Children To Write Sign Language Wikipedia Articles[edit]

We want to expand the grant to add an educational component. We have classrooms in three different cities in Brazil with Deaf children and adults who are skilled signers in Brazilian Sign Language (LIBRAS) and who already write their sign language in SignWriting mostly by hand. We would like to send Adam Frost, our Deaf ASL Wikipedia editor, down to Brazil to teach these classes about the Wikipedia using the new fonts. This will start activity in the Brazilian Sign Language Wikipedia: https://incubator.wikimedia.org/wiki/Wp/bzs . Are we allowed to add this educational component budget to this grant? We will need funds for Adam Frost's travel to Brazil, hotel, and some educational materials. --ValSutton (talk) 21:28, 24 March 2017 (UTC)

<grin> I would say that not only are you allowed, it is your grant request, it is a very strong argument in favour as well. </grin> Thanks, GerardM (talk) 06:40, 25 March 2017 (UTC)
@ValSutton: Thanks for this suggestion-- please do! It's great to hear that SignWriting is already an established component of these classrooms, and I think this kind of outreach work pairs well with the technical work you've already planned for this project. I JethroBT (WMF) (talk) 16:28, 29 March 2017 (UTC)
Excellent. We have received requests from 5 schools in Brazil that are interested in being included in this project. All of these school include training in SignWriting. Many others in Brazil are excited by this project and have been sharing the information around. We will finalize our list of schools to include, estimated the expenses, and add another section on the grant page. -Slevinski (talk) 18:21, 29 March 2017 (UTC)

Who will use this font and when?[edit]

Once completed, the 2-D font can be used immediately on any of the existing Wikimedia projects. In particular, the English Wiktionary has a large section for ASL. Recently, there was a discussion of this very topic about the need for full Unicode support and associated 2-D font.

Of particular interest for MediaWiki software is "Vertical writing support". A Phabricator task has been assigned to Brion about this particular issue. Now is a great time to coordinate our efforts.

The ASL Wikipedia on Incubator will make use of this font. Once the font is ready, we can convert the text and update the Gadget.

The 2-D font and associated characters will enable SignWriting text to be shared between Wikimedia projects and outside software. With everyone using the same standard, we can make significant progress with usability and coverage. Once SignWriting is easy to write and use in the MediaWiki environment (and elsewhere), the amount of written sign language will quickly expand.

This is a long term project. The written traditions for sign languages are very brief compared to other languages. Many sign languages have first generation writers. We believe that the Wikimedia projects will be a vital source of written sign language because we believe that a wiki style writing environment is a great way to create large quantifies of high quality writing.

Brazilian Sign Language, which has 3 million native speakers, is fortunate to have several generations of writers. SignWriting is required as part of the curriculum for several Brazilian Universities. More Universities are adopting SignWriting in Brazil every year. There are several regional groups who teach the reading and writing of Brazilian Sign Language. My favorite group is "Libras Escrita" which has over 30,000 likes on their Facebook page. Last year, a national conference was held in Brazil for the teaching of writing sign language. There are a thousand pictures of the event on facebook. If you look through the pictures, you will see people selling books and t-shirts covered in SignWriting. The momentum and excitement in Brazil is amazing to watch.

The Brazilian Sign Language Wikipedia will be started in the next year or two. I believe it will quickly surpass the ASL Wikipedia on Incubator and become the first sign language Wikipedia in production.

The vision statement of the Wikimedia Foundation:

Imagine a world in which every single human being can freely share in the sum of all knowledge. That's our commitment.

This vision is incomplete without written sign language. SignWriting is the way to write all of the various sign languages around the world. SignWriting has a proven track record of expanding access and improving technology. We do not just have a great idea and hope, we have a great set of tools and an international user base.

Valerie Sutton created SignWriting over 40 years ago. She gave her writing system to the world to freely use. Dozens of countries around the world have adopted her writing system.

In the past decade I've been involved with SignWriting, we've focused on developing free and open standards. The ASL Wikipedia has been one of our primary targets. It has been possible to write ASL articles for a while now, but the process and technology are complicated.

With our in-house software "SignPuddle", we created a place where writers from around the world could add their own signs. People have written over a million signs with our software. They have done this because they believe in literacy for sign languages.

There is an excellent paper titled A Survey of Those in the U.S. Deaf Community about Reading and Writing ASL. The results, 1/3 strongly support writing, 1/3 strongly oppose, and 1/3 see the benefit but aren't interested yet. I believe the participants might have been skewed towards those who support writing, but the conclusions are definite and obvious.

  • Chapter 6 conclusion: "The widespread use of a written form of ASL emerged as a factor that would persuade most of the respondents to adopt a written form of ASL . The majority of the respondents would want to learn to read ASL if reading materials existed. Also, most agreed that they would want to learn to read and write ASL if (1) their friends, (2) the local community, (3) the national Deaf community, (4) individuals they respected, or (5) Gallaudet University adopted a writing system for ASL. Therefore, if the majority of people in the Deaf community accepted literacy in ASL, then even the respondents who lack a personal motivation would learn to read and write ASL."

Our development has been an iterative process, where we build on the successes of the past and then build the next layer. We are nearing the completion of our technological foundation, but we are not done yet.

In the year 2010, we stabilized our symbol set with the "International SignWriting Alphabet 2010". This is the symbol set that was encoded in Unicode version 8.

In the year 2012, we finalized our character encoding model with "Formal SignWriting". We are currently using this character encoding for the ASL Wikipedia on Incubator. This character encoding is the basis for the supplemental Unicode proposal.

The 2-D font represent the next major step. This font will be compatible with the over 1 million signs written in SignPuddle and the 50+ articles of the ASL Wikipedia.

I can not stress enough the importance of this 2-D font and Unicode proposal. When dealing with an International audience, it's one thing to say we have this custom script that requires specialized software to use, and it's another thing to say we have an International script with a working font, 672 character accepted in Unicode and an addition 17 character proposed to finish the script.

The 2-D font and full Unicode strings will encourage many new people to get involved with SignWriting. I know this to be true because the lack of these has often been listed as the reason not to embrace SignWriting.

However, the 2-D font is not the final piece of the technological puzzle. The last piece is a variety of text editors that can read and write the full Unicode strings. There are 3 types of editors that will emerge: keyboard, drag & drop, and texting. Keyboarding has been a part of SignWriting for over 30 years. Drag & drop was introduced to SignWriting about 10 years ago. Texting requires about 6 taps to retrieve each sign from a dictionary. I have a texting prototype available and I'm sure others will have their own design.

Creating a SignWriting editor is complicated. The 2-D font will greatly simplify the text editor design and having full Unicode strings will encourage other developers to get involved.

With the Wikimedia Foundation's support of SignWriting, the Wikimedia projects will become a natural home for any of the world's sign languages. A small commitment today will reap large rewards for tomorrow and the foreseeable future. -Slevinski (talk) 21:06, 15 March 2017 (UTC)

Comments of Glrx[edit]

I am on the fence with this proposal. It sounds like a worthy project, but the more I dig, the queasier I get. I'm leaning decline because there should be more evidence that an ASL WP might succeed.

I agree with Blue Raspberry. The proposal hasn't described the task well, so it is difficult to estimate the work. What is involved with VOLT? There are Unicode character definitions, but they are somehow deficient and incomplete. There are no examples to give me context and no description of what the control characters do. I'll guess it is some sort of 2D kerning. I don't know whether the control characters are systematic or need extensive tweaking.

I'm also troubled with text editor problem. It seems there are many problems before an ASL WP would be ready for wide acceptance.

I JethroBT's question was about showing user interest in an ASL WP, but I take the answer as a bit off the mark. There are three contributors; others are in the wings if they could use phones and tablets. The answer seems too optimistic. I can look at other venues. People know how to use webcams; no written ASL required; no ASL input editor required. Does YouTube have many contributors of ASL content? Are there videos of people telling fairy tales in ASL? How about explaining math rate problems or gravity? I found some simple ASL math videos, but they had few views.

Unicode and text entry may be a problem, but give me some indication that things will take off.

The proposal is very push oriented: "if you build it, he will come." I have some trouble with that viewpoint. We're told that Sutton's system has been around for a long time, but its acceptance is not clear. Apparently no writing system has been officially adopted for ASL. Apparently people have not been publishing many books or pamphlets in written ASL. ASL seems to be a "spoken" language. I see political speeches that have ASL interpreters, but movies have CC English text rather than ASL in a corner window.

In a way, that makes sense. Deaf persons are immersed in their own national language environment, so proficiency in that national language is valuable. I get that its tougher to learn a language, but I'm not seeing evidence that introducing another written language, SignWriting, will make the task easier. The fundamental questions seem to be about improving deaf education rather than systems that will enable an ASL WP. What can be done to make things work better? That's a research in education issue.

Keogh's MA thesis is disheartening. Deaf children are not keeping up with their peers. After decades of trials, spoken ASL with written English has not levelled the playing field. Things need to get better, but the appropriate path is not clear.

I need to see more evidence that an ASL WP will take off. As it stands now, the proposal does not give much evidence of a demand for written ASL. Even if I assume written ASL would help language learners, I question whether significant topics would be covered in an ASL WP. Would somebody write an article on Einstein's theory of relativity? I don't see a strong end game for an ASL WP.

Glrx (talk) 21:39, 17 March 2017 (UTC)

Glrx, thanks for taking the time to review and consider the grant.
While the ASL Wikipedia is currently the most visible Wikimedia project using SignWriting, this project is really about improving Wikimedia projects in general for all sign languages. There are dozens of other sign languages used around the world, and Wikimedia has several other types of projects. While writing about Einstein's theory of relativity in the ASL Wikipedia may seem like a far fetched idea to some, consider writing new dictionary terms in an ASL Wiktionary. For ASL, new signs come and go because it is a very dynamic language that has a lot of regional variation. A centralized location of a wiki style dictionary would be a valuable tool.
I understand the concern that this is a push project; a project solving a problem without a definite demand, but there is already demonstrated demand within the English Wiktionary for writing ASL. They are currently using a custom ASCII solution, but would be interested in switching over to SignWriting if the font and full Unicode were available. Being able to use written sign language in any of the Wikimedia projects would be a huge improvement. A side benefit is that this project would promote diversity. There are dozens of sign languages around the world that could start Wikimedia projects. We are not building this for ASL only. From the Unicode Proposal for SignWriting:

SignWriting is currently being used to write the following Sign Languages: American Sign Language, Arabian Sign Languages, Australian Sign Language, Bolivian Sign Language, Brazilian Sign Language, British Sign Language, Catalan Sign Language, Colombian Sign Language, Czech Sign Language, Danish Sign Language, Dutch Sign Language, Ethiopian Sign Language, Finnish Sign Language, Flemish Sign Language, French-Belgian Sign Language, French Sign Language, German Sign Language, Greek Sign Language, Irish Sign Language, Italian Sign Language, Japanese Sign Language, Malawi Sign Language, Malaysian Sign Language, Maltese Sign Language, Mexican Sign Language, Nepalese Sign Language, New Zealand Sign Language, Nicaraguan Sign Language, Norwegian Sign Language, Peruvian Sign Language, Philippines Sign Language, Polish Sign Language, Portugese Sign Language, Québec Sign Language, South African Sign Language, Spanish Sign Language, Swedish Sign Language, Swiss Sign Language, Taiwanese Sign Language, Tunisian Sign Language.

Video is not appropriate for writing in general and definitely not appropriate for wiki style content and group editing. Imagine writing a 500 word English essay with only a voice recorder. Imagine someone else trying to change a few words of your voice recording.
Sutton SignWriting has been around over 40 years. This may seem like a long time in the tech world, but this is relatively short for a writing system. Before a writing system can be widespread, there is a huge amount of written and intellectual infrastructure that needs to be created and passed from one generation to the next. SignWriting is firmly established in Brazil with second and third generation writers. In Brazil, SignWriting has been adopted at several major Universities and there are several regional groups dedicated to teaching the writing of signs. From the English Wikipedia, the History of Writing:

In the early literate societies, as much as 600 years passed from the first inscriptions to the first coherent textual sources: i.e. from around 3200 to 2600 BC.

Regarding font development with VOLT, it will be an iterative process. I have already verified that our existing 1-dimensional fonts will import successfully into VOLT. There are extensive VOLT documents and design specs to review. The VOLT front end will be used to create specific examples. The VOLT exporter will be used to generate the VOLT Definition Language file for the specific example. The definition language will be analyzed and extended to the general case. A complete VOLT Definition Language will be created in an outside scripting language. This generalized Volt Definition Language will be imported into VOLT and then VOLT will create the OpenType Font file. The font file will then be tested against the project goals and then the process will repeat until the fonts are complete.
The existing 672 Unicode characters only cover the individual symbols, and do not cover internal structures or spatial layout. 17 new characters are proposed. These characters fall under 3 categories. 1) Additional symbol modifiers to fix sorting. 2) Number characters to represent distance. 3) Structural markers to identify the parts of each sign. These new characters will be used in every sign. The presentation SignWriting in Unicode Next contains more information.
The proposed Unicode strings are isomorphic with the Formal SignWriting strings. The Sutton SignWriting JavaScript Library has a function (ssw.unicode) for converting Formal SignWriting strings into the proposed Unicode strings. If you really want to understand how the Unicode strings are going to work, you may find it easier to review the ASCII Formal SignWriting strings first. The presentation SignWriting in an ASCII World is an excellent introduction.
The text editors are a complex topic. For keyboarding, each operating system has their own strategy, sometimes several different strategies. Physical keyboards are not standardized. There are probably around 100 different keyboards used around the world. Virtual keyboards vary across platforms as well. Fortunately, the design of the text editors is greatly simplified with a usable 2D font and full Unicode strings. -Slevinski (talk) 15:38, 18 March 2017 (UTC)
Thanks for your response, but it does not address my reservations.
I have no problem with the idea of a writing system for sign languages or efforts to improve it. The question in my mind is whether Wikimedia is an appropriate funder of such an effort. I have little experience with the topic. Thirty years ago I knew somebody who worked at the Walden School, but today I do not have a ready, independent, expert. A proposal directed to organizations who fund deaf-related or educational research or publish material related to signing would seem more to the point. They would have a better sense with which to evaluate the proposal.
I expect Sutton's system to cover multiple sign languages, but I don't see such coverage as being a deciding issue for WP. WP is interested in getting and supplying content from and for different groups. If a target group doesn't know how to read the language, then the content isn't very useful. If the target group doesn't know how to write the content, then the collection won't grow. Those are the issues that give me trouble when I look at the proposal.
I get caught up in some basics.
Say somebody is fluent at ASL. That does not mean the person can read or write with Sutton's system. How much effort does it take to become adept at Sutton's system? en:SignWriting#Advantages and disadvantages states, "The sheer size of its symbol set and the fine details which can be written create a challenge in learning how to write." That does not bode well for contributions to a WP.
If I were to look at some basic books about learning ASL, what would I see? To teach the word "goat", would there be a picture of a goat and a picture of someone signing goat? Instead of a person signing goat, would there be some SignWriting instead? How popular are SignWriting dictionaries?
I wonder about a parallel with en:International Phonetic Alphabet (IPA). I don't doubt the need for such a script or its utility; I've used IPA in some college courses where it was important to distinguish a t from a flap, and I've seen hundreds of sentences transcribed to IPA. It's a significant system for representing speech, but it is not used by many native language speakers.
When I was growing up, dictionaries would have their own pronunciation alphabets. IIRC, WP even used such an alphabet, but WP now gives IPA transcriptions. See, for example, wikt:colonel and wikt:indict; compare en:Colonel which does not give IPA. See also wikt:bird where IPA dialects are included. IPA is a reasonable practice, but WP also provides audio clips for pronunciations because not everybody understands the IPA.
There could be a similar effort with ASL. In addition to IPA, wikt:robin might have its ASL. Such an effort would be fabulous. But then how would that entry be represented? A video clip would be most accessible for everybody (and parallel the wikt audio clip). A script for ASL would also be good, but which ASL script? How many people would understand or benefit from a SignWriting transcription?
There are online ASL dictionaries. I can look up bird and find handspeak.com:bird. It has a video of a woman signing bird. It then has four frames selected as a printable version of the video sequence. Then there is "written ASL" for bird. The written ASL is an image (not a Unicode string), and it does not have the angular character of Sutton's system. It does not look like the "bird" entry on page 57 of Sutton's American Sign Language Picture Dictionary.
If I start pulling on threads, I get en:Stokoe notation (started for his dictionary), en:ASL-phabet, en:HamNoSys, en:SignWriting (Sutton), and en:si5s. http:www.ASLwrite.com is a fork of si5s. I get the sense that the choice of ASL writing systems is very unsettled. Where is a metric that tells me the popularity or acceptance of each script? I'm a content contributor to WP, so I'm also conscious about the notion of reliable secondary sources. The en:SignWriting article has many primary sources such as theses and journal articles, but it is missing secondary sources. I cannot trust The SignWriting Journal to be an unbiased source. Where is a reliable, independent, secondary source that surveys the various ASL writing systems and rates their importance? The WP article has an interesting statement about published SignWriting in college newsletters and YouTube captions, but the statement has a {{Citation needed}} tag. I'm scratching my head about how many subscribers such a newsletter would have (20?) and how many colleges would have such newsletters. (A google search for "SignWriting in Unicode Next" does not give me a feeling of acceptance or widespread interest.)
That takes me back to the WP community is not the best one to evaluate the merits of this proposal and not the best to fund it. If there were evidence of strong acceptance on one ASL script, then the situation may be different. I'm OK with "push"ing a good idea, but this proposal toubles me.
In my initial comments, I brought up videos not as a method of writing but as a measure of communication volume in ASL. I'm looking for a measure of how much exposition is published in ASL, and my sense is not much. I loved the ASL jokes in Children of a Lesser God (the play, not the 1986 movie); I loved the sound of the Latin oration at Harvard's commencement. Consequently, I would expect to find ASL speeches and other contributions in a video collection. Dr. Janice Smith-Warshaw delivers the Gettysburg Address using American Sign Language (ASL) has only 3000 views. If there is little exposition in ASL, then the prospects for community contributions in written ASL is low. I'm unstudied on the subject, but my guess is the en.deaf community speaks ASL but writes English. The proposal cited an MA thesis, but the interest of that thesis was performance at grade level in mainstream tasks such as math and written English.
The proposal may be a great idea, but it does not persuade about its impact or reach. It seems there is a double push involved: one push is to accept SignWriting as the ASL script, and the second push is to get the deaf community to start publishing significant content in ASL. I'm dubious about both pushes. The proposal is too much about the technical merits of SignWriting and too little about its impact. If this proposal were about including 5,000 ASL signing videos in en.wikt, then I'd recommend it, but the proposal is something different. The proposal promises Unicode SignWriting, but it doesn't persuade that SignWriting will benefit a substantial percentage of the deaf community or a competent SignWriting Unicode representation will encourage others to insert 5000 ASL entries in en.wikt or some other WM venue. Maybe it will happen, but I want to see the arguments for coverage and skill.
see also handspeak.com:Sign Language Writing and Calligraphy
Glrx (talk) 20:05, 18 March 2017 (UTC)
Dr Maria Galea’s Ph D thesis “SignWriting (SW) of Maltese Sign Language (LSM) and its development into an orthography: Linguistic considerations” includes the following statement.

“Additionally during this work, it was a challenge to keep up with the influx of new technologies for SW and other works related to SW. This proves furthermore that SW is becoming widespread, uncontainable and untraceable. In the same way that works written in and about a well developed writing system such as the Latin script, the time has arrived where SW is so widespread, that it is impossible in the same way to list all works that have been produced using this writing system and that have been written about this writing system.“

Deafness as been called the invisible disability. Just by looking at someone, you can not tell if they are deaf. In our modern age, this saying has a second aspect in that just because you can't find something about deafness online doesn't mean it doesn't exist.
It is a false idea that the deaf will just use the spoken language of the area where they live. Some deaf are very skilled at reading and writing the dominant spoken language, others have only a limited understanding and others have none. Stefan Woehrmann is a teacher in Germany who teaches the deaf that have been failed by the mainstream system. His deaf student have little to no understanding of German. He uses SignWriting to teach them to read and write their sign language. He then uses their literacy in their primary language to teach them their second language.
Glrx, I am not able to provide the assurances you desire. There are a lot of other points I could address, but they would not address your primary concern. -Slevinski (talk) 14:31, 19 March 2017 (UTC)
I follow the link to pull down Galea's thesis (another primary source), and it wants me to login. So I search for the title and find a copy on the SignWriting website. I start reading it. There's significant funding by the European Union European Social Fund (ESF), and that's what I want to see. Then I read:
I also thank my co-supervisor and the inventor of SignWriting, Valerie Sutton. Valerie has been a crucial provider of information concerning SignWriting and its roots. She has been of constant support, not just during this work, but right from the very beginning of my use of SignWriting 16 years ago. SignWriting is 40 years old this year, 2014 and I am so happy that this work has been produced during the 40th anniversary of Valerie’s invention.
Galea is not independent of Sutton.
There's also an acknowledgment for Slevinski and SignPuddle.
I want some independent sources telling me SW is a good idea. There must be journals and annual surveys of progress and goals.
I would expect those involved with SW to able to point to statistics that address the use of various forms of written sign language. What percentage of the deaf population can read which form?
Galea's thesis states deaf-at-birth incidence is 0.01% and estimates the deaf population of Malta at 400. Ten of those 400 have had formal training in SW; they may have passed some of their SW skills on to others. There may be another 3 SW-able persons because the sample size of the survey was 10 and only included 7 of the formally trained. After 3 days of training, LSM speakers "were able to write a small piece of text" using SW. (p 16.)
Those numbers scare me. It says SW is accessible to 10/400 = 2.5% of the Maltese deaf population; The percentage may only be that high because there was a decided effort to train them. If I take the US population to be 300M, then there are 30,000 deaf at birth. Taking 2.5% of those gives me a US population of less than 1000 who are capable of SW.
Galea's thesis also does not give me a broad impression about applicability. Cf 112 is not helpful. Vague SP stats at 315–317. Malta SW shows slow growth p 318. SW is appropriate for academic description of signed languages. SW is used in dictionaries. SW is used in training materials. The Bible has been expressed in SW (the Bible seems to be the dominant ASL SW activity: 13K entries compared to 1K US pp 319–320).
I don't doubt that Unicode SW has merit and utility. I don't doubt that Slevinski has the knowledge and skills for the task. I don't doubt difficult challenges for the deaf community. I'm not sure about WM's potential audience.
Glrx (talk) 20:13, 20 March 2017 (UTC)

Random notes[edit]

Looking for high hit rates on signed videos.

One editor did most of the articles on WP incubator.

Search for ase-Sgnw langtag dimmed quickly.

Did not find IETF script tags for HamNoSys or si5s. Glrx (talk) 20:06, 21 March 2017 (UTC)

Sign Videos Beyond ASL[edit]

If you looking for high hit rate signed videos, you'll want to expand beyond just ASL. Here are a few videos using Brazilian Sign Language.

Here are a few videos about Brazilian Sign Language and SignWriting.

-Slevinski (talk) 21:35, 21 March 2017 (UTC)

Position[edit]

I would regretfully decline this proposal. That conclusion makes me feel awful, like I'm stepping on someone who is already down. The proposers are well intentioned and altruistic, but at this time the proposal doesn't gel.

I think the proposers could quickly cite the acceptance percentages for various ASL writing systems, but they have not. Instead they point to claimed success in Brazil and Malta, but they give qualitative rather than quantitative descriptions. Brazil's population is 200 million and would have 20,000 deaf at birth. What reach does SW currently have in Brazil? Are we at Malta's 2.5 percent? The proposers should know those details; why aren't they stated.

I'm also confused by the mission/purpose of SW. It unquestionably serves a research purpose akin to the IPA. At times, SW's utility seems to be as an aid in teaching sign language: printed dictionaries, workbooks, and short stories. That utility is still experimental: proponents hope it will help teaching, but there are no solid results yet. Perhaps there's a Wikibooks angle for publishing such material, but it's a one-way audience.

SW utility outside teaching is sketchier. Once we leave teaching, is there a significant sign language only publishing segment? A significant interest in publishing such materials? I'm not seeing high publishing volume. I love Sign Duo and Matlin, but I want to see more. For comparison, I can search YouTube for "Navajo songs" and find many high volume videos. The Navajo population is 170,000 (roughly 10 times US deaf at birth), so a Navajo song that gets 500,000 hits compares to an ASL video at 50,000 if the mechanism is SN1 (one who wants to publish just publishes). Sign Duo and Matlin are not targeting an ASL-only audience.

My impression of SignPuddle is that it attempted to be a Wikipedia for multilingual SignWriting: users would contribute articles for everybody to read. From Galea's thesis above, the effort didn't take off because the contributions have been small. The Brazilian SignPuddle had 143 literature entries in 2014 (up from 132 in 2012). (Galea p. 319.) That is not high growth. Statements that a bsl.WP will "naturally grow" sound nice, but it is inconsistent with the SignPuddle experience. I think the proposers have lots of data from their SignPuddle experiences. That SignPuddle data is neither presented nor explained.

I'm not following the ASL WP, and I do not know the metrics used to judge its merits. It is odd that the SW's Adam Frost had only contributed one article to asl.WP by October 2016 even though Frost is apparently referred to as "the ASL Editor of the ASL Wikipedia". Nancy Romero is more deserving of the title with 49 articles; second in line is Jason Nesmith with just 4 articles.

The teaching trips to Brazil initially struck me as a good idea, but the trips are premature. The effort is about teaching Sign Writing using the new 2D font development. However, a Unicode SW input editor isn't available yet; my sense is it is left for later: "However, the 2-D font is not the final piece of the technological puzzle. The last piece is a variety of text editors that can read and write the full Unicode strings." It's also odd that Frost "will fly to Brazil to 5 well-known Brazilian schools where SignWriting is an established part of Deaf Education and Sign Language Education". The 200 targets of the trip are already exposed to SW, so the goal would be computer interaction. The better approach is to finish the 2D font, develop a reasonable editor(s), and then seek a trip to Brazil to teach the editor. I'm also puzzled because Frost knows ASL, but we are not told if he also knows LIBRAS/BSL. How similar are ASL and BSL? (They may have similar French Creole roots.) Would Frost make a good instructor for a BSL audience?

I think the goal of proposers is stated in a quote from a January revision: "With the creation of the font and standardization of the new characters, Sutton SignWriting will be firmly established for all time." That may be an admirable goal, but I'm not seeing independent secondary sources that tell me the goal has merit. Hasn't the NIH / NIDCD commissioned some report that addresses SW?

The NIDCD page is depressing; it suggests that many hearing problems are not detected until it's too late for good language development.

The WMF needs independent expert advice about this topic, and that advice is not in the proposal.

Glrx (talk) 00:02, 4 April 2017 (UTC)

@Glrx: Ethnologue estimates 3,000,000 native BSL signers, not 20,000. --Yair rand (talk) 22:03, 5 April 2017 (UTC)
The claim is dubious. Ethnologue states BSL is a first language for 3,000,000, but the statistic is probably a mistake that should say BSL is a second language of 3,000,000. The large number is a result of en:Legal recognition of sign languages#Brazil and reflects BSL training rather than BSL as a first language. The 20,000 figure I used results from Galea's population statistic and would be potential primary language BSL. I don't want to detract from Brazil's efforts, but those 3,000,000 probably have limited ability in BSL and primarily speak and write in Portuguese. The level of BSL training/fluency is not clear. Furthermore, BSL law suggests that the Brazilian government would be open to SW proposals (such as this one), would be in a much better position to evaluate them, and would have no trouble funding proposals at a penny a speaker. The Brazilian government would have ready access to sign language experts who could offer significant evaluations; WMF has few, if any, sign language experts. FENEIS would be another organization. The group-think of SW insiders is not persuasive. I don't like being skeptical/negative, but I'm not seeing good, outside, recommendations that such a script fits WMF's mission (I do not speak for WMF). Glrx (talk) 23:54, 5 April 2017 (UTC)
I'm not sure why you think the statistic is a mistake. The census shows a deaf population of 5750800. --Yair rand (talk) 00:37, 6 April 2017 (UTC)
That makes the proposal success prospects dimmer.
What do you mean by "deaf"?[1]
The US has a huge portion of its population with some form of hearing loss.[2]
The CDC states that many people have detectable hearing loss.[3]
I used the statistic that Galea stated: 0.01% of the population is deaf at birth. (Galea p 14.) The proposers pointed me to her thesis, and I got the thesis off the SW website. On that page, she states:
The only statistics relating to hearing loss that are available are from the KNPD13 (National Commission for Persons with Disability, Malta) who provides a yearly report on disability in Malta. This report includes annual statistics on hearing-loss. The number of Maltese people with hearing-loss amounts to 834 in 2012 (p. 43). This number includes all types of hearing loss, and not only profound deafness at birth which results in the inability to acquire speech naturally through hearing. As Johnston (2004) states: “it is only children with an early and profound hearing loss, and many with an early severe hearing loss, who are likely to be lifelong users of sign language.” (p. 358)
If we take a CDC figure of 0.1% instead, then the statistics are 10 times worse for the current proposal. If the signing deaf population of the US were in the 300,000 range, then I would expect YouTube videos comparable to Navajo videos.
If we take your figure of 1.5% deaf (4,500,000 ASL signers), then where is the ASL literature/storytelling? The ASL videos should have exploded.
Glrx (talk) 03:25, 6 April 2017 (UTC)
Also, the abstract for "SignWriting in Brazilian Deaf Education: 1996 to present" that was given at the 2014 SW symposium has some relevant info on SignWriting in Brazil. --Yair rand (talk) 22:31, 5 April 2017 (UTC)
Such publications were already addressed as a dependent journal/symposium. Furthermore, Santa Catarina is not an unbiased institution with respect to SW; see en:Brazilian Sign Language#Alphabet. Glrx (talk) 23:54, 5 April 2017 (UTC)

Another issue that has been in the back of my mind is the ringing smart phone in the Sign Duo video and some comments about software for phones above. Texting is infectious, and I'd expect it to be used by the deaf. Here's an article from 2010:

The article also raises video chatting. (My experience is people would much rather send a text rather than make a phone call or a video conference.) Glrx (talk) 20:54, 6 April 2017 (UTC)

Another issue in the back of my mind is the writing rate and areal density. People speak at about 120 words per minute. My sense is that signers have a similar rate. People type at 40 words per minute. I wonder how fast SW is written.

Texting with SignWriting is in development. Using a dictionary based approach, most any sign can be selected with 4 taps. A prototype is available in SignMaker. On the top right, click the "Click Search" button. Then start clicking on symbols on the right to limit the search results on the left. SignMaker is only intended to work on one sign at a time, but the design is solid. This design will be integrated into a texting app in the future.
Writing by hand is different than writing on a computer. In 1982, sign language stenographers where able to write sign language at normal signing speed. For general hand writing, the speed is comparable to English. The computerized model is called block printing. When written by hand, a type of cursive is often used that is less detailed than the block printing. I have a short article about the subject available if you want more information.
On computers, the keyboarding interface and the drag&drop interface is much slower than touch-typing with English, but with dictionary integration the speed is much improved. -Slevinski (talk) 21:23, 6 April 2017 (UTC)
Thanks. Glrx (talk) 22:59, 6 April 2017 (UTC)

Who are we and what are we there for[edit]

The Wikimedia Foundation has as its goal "share in the sum of all knowledge". What we do is writing Wikipedia articles, Wikisources, Wikidata items you name it. We do this in any and all languages. The requirement is simple. Any language, any public may share in this endeavour. SignWriting is a script. It enables all the sign languages in the world to be written. It is not limited to the American Sign Language, SignWriting is used for a multitude of languages around the world, Tunesian, Saudi Arabian, Brazilian.. (just a few).

The point of the SignWriting grant is that finally there will be a font that will enable all the sign languages in the world to be written. It will be possible to write it using Unicode, the standard for all written text on the Internet. It will enable.

As a member of the language committee I can attest that all sign languages are eligible (the ones that have an ISO 639-3 code). I have been in contact with people from the SignWriting community for a long time and their commitment is huge. Yes, they are committed to the children they teach. Ask yourself what the difference with a language you can write and one that you cannot write.. Yes, children will be asked to write articles. Sign Writing will be huge and it will be worth every cent we invest because we enable and thereby open up a whole new group of people that do not even have school books in their own language.. That is what it will do and all we do is enable them. Thanks, GerardM (talk) 05:56, 21 March 2017 (UTC)

Eligibility confirmed, round 1 2017[edit]

IEG review.png

This Project Grants proposal is under review!

We've confirmed your proposal is eligible for round 1 2017 review. Please feel free to ask questions and make changes to this proposal as discussions continue during the community comments period, through the end of 4 April 2017.

The committee's formal review for round 1 2017 begins on 5 April 2017, and grants will be announced 19 May. See the schedule for more details.

Questions? Contact us.

--Marti (WMF) (talk) 19:53, 27 March 2017 (UTC)

Dear Slevinski,

Thank you for submitting this proposal.

I have a question for you: Can you clarify whether the work you propose is unique to Wikimedia project, or whether your intent is to develop software that would be useful in any context, not just Wikimedia projects? So long as your work will, in fact, benefit our projects, it is eligible for funding either way. However, because your budget is large and this round looks like it will be particularly competitive, it would be helpful to know which pieces are Wikimedia-specific and which are more global.

Kind regards, --Marti (WMF) (talk) 02:16, 3 April 2017 (UTC)

Hi Marti (WMF),
The majority of the grant covers the font development and the Unicode proposal. These elements are universally applicable, and they are not Wikimedia specific. While the major portion of the grant isn't Wikimedia project specific, we believe that the Wikimedia projects will be one of the greatest beneficiaries of this grant. There are 5 main benefits that will improve the Wikimedia projects.
1) Unicode rather than ASCII
The ASL Wikipedia on Incubator uses a lite ASCII markup called Formal SignWriting to encode the script, because full Unicode isn't available. Even without Unicode, some people involved with the Wikimedia projects have adopted SignWriting (see Yair Rand's SignWriting Keyboard and Translate Wiki). There is real interest in SignWriting and a full Unicode design will encourage additional people to get involved with SignWriting development (see ASL section of English Wiktionary).
2) Simplify development
With a 2-D font, SignWriting can be used with any of the Wikimedia projects without having to handle the text in a different way than every other script. The 2-D font and full Unicode reduces the complexity of dealing with SignWriting and promotes universally accepted standards.
3) Integrate Vertical Writing Mode
There is currently an effort to enable Vertical Writing Mode in MediaWiki software. The creation of the 2-D font will enable development to move forward with real-world examples in all of the various vertical scripts. Without the 2-D font, the vertical writing mode efforts will not cover all of the use cases and SignWriting integration will need to happen at a later time.
4) Easy to share
A great benefit of the 2-D font and full Unicode is that it will be easy to share information with other projects. Users will be able to copy SignWriting from the Wikimedia projects and use it elsewhere, and users will be able to paste SignWriting into Wikimedia project from elsewhere.
5) By the Deaf, for the Deaf (and others)
We've found a common theme for all of the sign language Wikipedia projects: for ASL, for Tunisian Sign Language, and now for Brazilian Sign Language. There is a pride and dignity associated with the idea of a sign language Wikipedia. The idea that the Deaf will be able to personally build up a valuable resource with their own efforts is empowering. Often times, the Deaf have had others decide what was in their best interest and then develop/impose outsiders' ideas on the Deaf. That is the beauty of SignWriting and Wikipedia: we are enabling the Deaf rather than providing the Deaf with what we think they should have. With SignWriting, we don't tell them how to write a sign, but give them the tools so that they can write a sign how they see fit. With a sign language Wikipedia, we're not imposing our ideas on them, but giving them the resources they need so they can develop it themselves, in whichever way they choose to develop it.
The SignWriting community has already accomplished a great deal and there is a tremendous amount of interest in sign language Wikimedia projects. Now is the time to capture this enthusiasm, coordinate our efforts, and build upon it for the future. I believe the Brazilian Sign Language Wikipedia will become a great success and it will encourage other sign languages to start their own Wikipedias. This grant is a significant investment that will pay dividends for years to come. Imagine a world in which every single human being (including the Deaf) can freely share in the sum of all knowledge. -Slevinski (talk) 22:03, 3 April 2017 (UTC)

Comments[edit]

It is an interesting proposal although I share some concerns of Glrx. So, some comments, questions and suggestions:

  1. The proposal looks like a mechanical sum of two unrelated proposals: the font development and student training. So, I think that project should be split in two.
  2. I do not think it will be possible to use newly developed font (even if successful) to teach students as its use requires extensive software support, which will not be promptly available. Am I right?
  3. As I understand 17 control characters that are necessary for the signwriting are intentionally not part of Unicode. How the signwriting with your font is going to be implemented in this case?
  4. I suggest that you spell out the proprietary software that you want to purchase. And also can you name your consultants?
  5. Why did you choose VOLT as the main font software?

Ruslik (talk) 19:31, 8 April 2017 (UTC)

1) The initial grant proposal for only for the font development. The student training was added to show the interest in sign language Wikipedias. In the last 2 weeks, we have been contacted by 21 Brazilian schools that are interested in a Brazilian Sign Language Wikipedia.
2) The student training is scheduled for late 2018. The plan is to make a prototype font available as quickly as possible so that software support of the fonts can begin immediately. I'm reasonably certain that the fonts can be perfected and the software support can be implemented before the student training begins. The student training will also provide usability information that can be used to improve the software.
3) The font will use the 672 approved characters along with the 17 proposed characters. The 17 new characters have not yet been officially proposed to Unicode. Once the new font has been created, an official proposal will be submitted to the Unicode Technical Committee for the 17 new characters with the new font playing a vital role in the proposal.
4) There is a variety of font software available and I'm not sure what I'll need. I may be able to use free software entirely, but I do not know yet. I can create a list of the software I've considered. For the consultants, I do not have a list of people picked out yet. With the font development, there are a variety of unusual design issues I will be handling. If I am having a particularly difficult time with any of the specific issues, a consultant may be able to provide insight or targeted development to help me overcome the particular issue.
5) When I attended the Unicode Technical Committee (UTC) meeting #148 in August of 2016, I had the opportunity to sit with Andrew Glass for a few hours. Andrew created the design for the Universal Shaping Engine (USE). All scripts require a font and an associated shaping engine to properly display text. The benefits of the Universal Shaping Engine is that it is generic so that scripts only need to create a font that targets the Universal Shaping Engine and therefore the script will not need to create a special shaping engine. Andrew walked me through VOLT and discussed the Universal Shaping Engine. He showed me his working font for complex hieroglyphics. He was certain that VOLT was sufficient for the task of created a 2-D font for SignWriting. The font will be created with VOLT and the font will leverage the Universal Shaping Engine. As far as I know, VOLT is the only font software that can be used to design specifically for the Universal Shaping Engine. The only other alternative to VOLT/USE is to develop for Graphite. We already have a proof of concept font for Graphite, but it isn't production ready and Graphite is limited in adoption.
-Slevinski (talk) 20:44, 8 April 2017 (UTC)