Grants talk:IEG/Alt text tools

From Meta, a Wikimedia project coordination wiki

SEO lessons[edit]

Google encourages alt text by giving higher rank to pages which use it (at least for the keywords), hence all SEO workers are hard at work adding it. Perhaps we should find ways to make the army of SEO/PR editors work on alt text too. ;-) --Nemo 08:35, 16 September 2015 (UTC)[reply]

9/29/15 Proposal Deadline: Reminder to change status to 'proposed'[edit]

Hi Dispenser,

The deadline for IEG submission this round is September 29, 2015. To submit your proposal, you must change the status from "draft" to "proposed." I see that your draft for this ideas still has a number of empty fields. If you have any questions or would like to discuss your proposal, let me know. We're hosting a few IEG proposal help sessions this month in Google Hangouts. I'm also happy to set up an individual session. Warm regards, --Marti (WMF) (talk) 20:42, 20 September 2015 (UTC)[reply]

Getting feedback from disabled users[edit]

At Wikimania a couple years ago we had a workshop hosted by a blind Wikipedia editor (can't remember their username). One of the most surprising things that I learned at that workshop was that they hated images with alt text. The reason was that almost all images on Wikipedia already have captions and the alt text tended to be mostly (or completely) redundant with the captions. Thus they would have to listen to two descriptions for each image when using a screen reader. For them, this was annoying. I'm not sure if a blind Wikipedia reader (as opposed to an editor) would feel the same, but I think it would be important to survey some actual disabled users before moving ahead with this project. Kaldari (talk) 23:50, 29 September 2015 (UTC)[reply]

You are probably referring to user Graham87 ? Writing good alt text is certainly trickier than it seems. And I believe that yes most alt text that editors write are not good, or are redundant. Dodoïste (talk) 14:01, 30 September 2015 (UTC)[reply]
While the blind will benefit most from text alternatives we should not limit it to them. There are other mediums such as text browsers, TTS audio books, and braille books. In the past we've asked (and forced) article authors to write text alternatives and this is the reason quality varies wildly, especially as the guidelines changed. This proposal attempts to create a niche editor who specializes in creating text alternatives. Additionally, the co-grantee has years of experience in the field. Dispenser (talk) 16:12, 30 September 2015 (UTC)[reply]
Ok, that sounds great.
I have to point out that our current guidelines on alt text are also meant to compensate for the imperfect way MediaWiki handles linked images (most images on Wikipedia). Most images in articles are purely decorative and do not carry information needed to understand the article. Ideally those images should have blank alt text. But if we leave the alt text blank on a linked image, the filename is read aloud by the screen reader which produce a horrible result. So we have to provide minimally annoying alt text like "|alt=photograph" or "|alt=refer to caption" when the caption provides enough information. This is a burden on editors, I believe the software could do this automatically. The editors should only focus on images that carries important information that can not be written in the caption.
Since you're going to tackle this giant task, I figured you might as well improve the software to make the task easier. :-) Cheers, Dodoïste (talk) 21:32, 30 September 2015 (UTC)[reply]

Eligibility confirmed, round 2 2015[edit]

This Individual Engagement Grant proposal is under review!

We've confirmed your proposal is eligible for round 2 2015 review. Please feel free to ask questions and make changes to this proposal as discussions continue during this community comments period.

The committee's formal review for round 2 2015 begins on 20 October 2015, and grants will be announced in December. See the schedule for more details.

Questions? Contact us.

Marti (WMF) (talk) 05:31, 4 October 2015 (UTC)[reply]

Room for automation?[edit]

There are probably millions of images missing alternative text in Wikimedia projects. In this context, 60 more articles per month are of course welcome, but still far from changing the trend. Is there room for automation in your project? At least to avoid users getting the filename, which is worse than nothing most of times. You mention that your tool will make it easy and fun for editors, but no matter how, doing that work manually will take a lot of time that perhaps could be invested in editorial work less likely to be fulfilled by automated tools.--Qgil-WMF (talk) 05:44, 5 October 2015 (UTC)[reply]

So I thought about it. The way to automate it is to mash up a description from categories; those categories could be generated by a deep learning algorithm trained on images already categorized. However, Commons categories are a mess which need their own IEG to fix, the text alternatives would lack nuance, and Wikimedia is unlikely to invest $1 million in an artificial intelligence team (evidence: The repeated NOs to my 24 TB requests for fact prover).
Some people have suggested "default alt text" as a magic bullet. The bulk of images (enwiki: 2,919,153) are only used once. And copying text alternatives may not necessarily be appropriate in another context. Such as using bubbles in a soda to illustrate carbonated beverages and then focal planes.
Finally, we're all volunteers (well most of us) adding work does not mean work is abandoned in another area. In fact, it may attract people who cannot do the bantered editing tasks, but still would like to contribute. Dispenser (talk) 03:54, 7 October 2015 (UTC)[reply]
There is no room for automation when it comes to alt text. Nope. But the MediaWiki software could be improve and do 90% of the job. See my comment above. Dodoïste (talk) 19:13, 7 October 2015 (UTC)[reply]

How will the tool work?[edit]

You mention that using this tool will be easy and fun. Would it be possible to describe the workflow, to get an idea of your plans?--Qgil-WMF (talk) 05:45, 5 October 2015 (UTC)[reply]

--Hi Qgil-WMF - The tool will be based off of this working prototype. Our hope is to expand this tool to be more engaging, informative, and fun through the creation of documentation, refining the visuals, and actively marketing the tool through edit-a-thons and meetups. --Arthur.frick (talk) 19:50, 31 May 2016 (UTC)[reply]

Questions from Superzerocool[edit]

Hi, thanks for your proposal. As I read, it seems clear and feasible the project, but I have some thoughts about the next topics:

Thanks in advance. Superzerocool (talk) 12:23, 16 October 2015 (UTC)[reply]

Aggregated feedback from the committee for Alt text tools[edit]

Scoring criteria (see the rubric for background) Score
1=weak alignment 10=strong alignment
(A) Impact potential
  • Does it fit with Wikimedia's strategic priorities?
  • Does it have potential for online impact?
  • Can it be sustained, scaled, or adapted elsewhere after the grant ends?
(B) Innovation and learning
  • Does it take an Innovative approach to solving a key problem?
  • Is the potential impact greater than the risks?
  • Can we measure success?
(C) Ability to execute
  • Can the scope be accomplished in 6 months?
  • How realistic/efficient is the budget?
  • Do the participants have the necessary skills/experience?
(D) Community engagement
  • Does it have a specific target community and plan to engage it often?
  • Does it have community support?
  • Does it support diversity?
Comments from the committee:
  • Supports our need to make our information accessible to people with limited vision.
  • I like the game idea with barnstars and using the dabsolver approach to attract micro contributions.
  • This gadget may conflict with the highly popular pop-ups gadget, making me wonder if the greater wiki community will be able to see the effects of their edits. Without a clear and present need, it will be a hard sell to contributors to get them to add alt text.
  • The presentation and the idea needs further development before funding.
  • Budget needs explanation (This has been updated --Arthur.frick (talk) 19:49, 31 May 2016 (UTC)).[reply]
  • Not much support, but this may be because we serve the community with limited vision so poorly. Currently, there is only a machine-voice over method for audio articles because any spoken version is by definition historical and quickly becomes outdated. Audio versions are rarely if ever updated.
  • Lacks community involvement (the sections have been updated to reflect meetups and edit-a-thons --Arthur.frick (talk) 19:49, 31 May 2016 (UTC)).[reply]
  • I would like to see a mockup with screen shots of results. That may be what that link is supposed to be, but I see the same pop-up on mouse-over for all of them (probably because I have the pop-up gadget enabled). (Please view the Working Prototype --Arthur.frick (talk) 19:49, 31 May 2016 (UTC))[reply]
  • The project is interesting and can help some people (like blind people) for whom alt attribute is fundamental but I don't see a well developed project plan. The budget needs an explanation of how it was calculated. (This has been updated --Arthur.frick (talk) 19:49, 31 May 2016 (UTC))[reply]
  • I'd like to see a response to the detailed budget question and Quim's request for a description of what they're actually thinking about. (This has been updated)

Round 2 2015 decision[edit]

This project has not been selected for an Individual Engagement Grant at this time.

We love that you took the chance to creatively improve the Wikimedia movement. The committee has reviewed this proposal and not recommended it for funding, but we hope you'll continue to engage in the program. Please drop by the IdeaLab to share and refine future ideas!

Comments regarding this decision:
While this project could assist the blind for whom alt attribute is fundamental, the proposed tool may conflict with the highly popular pop-ups gadget. The committee would like you to explore how the greater wiki community would be able to see the effects of Alt text. We hope to see this idea develop in future and would be happy to have you return to IEG as your plans develop in response to the feedback shared during this process!

Next steps:

  1. Review the feedback provided on your proposal and to ask for any clarifications you need using this talk page.
  2. Visit the IdeaLab to continue developing this idea and share any new ideas you may have.
  3. To reapply with this project in the future, please make updates based on the feedback provided in this round before resubmitting it for review in a new round.
  4. Check the schedule for the next open call to submit proposals - we look forward to helping you apply for a grant in a future round.
Questions? Contact us.
The objections in the previous section made some sense, but «conflict with the highly popular pop-ups gadget», what? Nemo 23:12, 4 December 2015 (UTC)[reply]

Eligibility confirmed[edit]

This Individual Engagement Grant proposal is under review!

We've confirmed your proposal is eligible for review and scoring. Please feel free to ask questions and make changes to this proposal as discussions continue during this community comments period (through 2 May 2016).

The committee's formal review begins on 3 May 2016, and grants will be announced 17 June 2016. See the round 1 2016 schedule for more details.

Questions? Contact us at iegrants(_AT_)wikimedia · org .

--Marti (WMF) (talk) 04:46, 28 April 2016 (UTC)[reply]

Aggregated feedback from the committee for Alt text tools[edit]

Scoring rubric Score
(A) Impact potential
  • Does it have the potential to increase gender diversity in Wikimedia projects, either in terms of content, contributors, or both?
  • Does it have the potential for online impact?
  • Can it be sustained, scaled, or adapted elsewhere after the grant ends?
(B) Community engagement
  • Does it have a specific target community and plan to engage it often?
  • Does it have community support?
(C) Ability to execute
  • Can the scope be accomplished in the proposed timeframe?
  • Is the budget realistic/efficient ?
  • Do the participants have the necessary skills/experience?
(D) Measures of success
  • Are there both quantitative and qualitative measures of success?
  • Are they realistic?
  • Can they be measured?
Additional comments from the Committee:
  • Nice idea, but needs to work with the Wikidata "depicts" property, which would help prioritize images, as currently only significant images are on Wikidata.
  • Important work that fits with Wikimedia’s strategic priorities but has limited evidence that the project can be sustained after the proposed period. There is also not enough details about volunteer engagement, which in my opinion would compromise the potential for online impact.
  • Alt is fairly important to the disabled, which is a big plus here. But the approach is quite hard. Why build a wheel to fix a hole without fixing the leak first? That is the uploading tool aspect of the project. Many photos would be updated at some point, and providing/generating an Alt to DB for displaying it later sounds stronger than a leaderboard.
  • This is a massive problem that has remained neglected due to the overwhelming effort needed to make a dent in the backlog.
  • Yes, looking at articles from the perspective of the visually impaired is always a good thing. Unfortunately this is too context-sensitive to even use the same alt-text for other Wikipedia uses of the same image.
  • I like the approach of a Dab solver-like tool/leaderboard plus hub/project page. A measure of success is provided (60 more articles per month to have text alternatives added for their images) but no plan for how to get there. Having a tool won't on its own lead to the contributions or impact.
  • This project seems like a game, and could be fun for some.
  • There is plenty of research that can help with this project. Such a tool would be helpful in many ways. I would also suggest a focus on populating file descriptions on commons with image captions.
  • The impact may be minimal as the project may generate reams of text that would still need to be curated and managed as any other part of an article.
  • More details about budget would be nice, as requested in the last IEG round.
  • The applicants have strong backgrounds, but have underestimated for success in the long-run.
  • There is no way to judge one of the applicant’s qualifications as he is unknown to the community.
  • Seems to have support, but the number of visually impaired readers is minimal compared to sighted readers.
  • I see some endorsements but no target communities.
  • English Wikipedia is the target which is the main wiki for authors. It's easier to migrate to another wiki projects.
  • There isn't a developed plan for community engagement in this proposal at the moment. I would suggest using Wiki labels to gather feedback from users in order to train machine learning classifiers. For example, you could give users images and caption and they can rate the caption as "good" or "bad."
  • I would like more clarification on budget.
  • This tool could improve the equal access to knowledge for blind people. There are some improvements from the previous round, but still has weak support. I will stay neutral.
  • If this proposal included use of the ‘depicts’ property (P180) in Wikidata, I would vote for it.
  • We saw this project in the last round and unfortunately it looks like not much has changed: the budget is not explained and I see no further plans for community engagement, though perhaps work on a mockup has progressed since December. I also note from some of the comments on the talk page and elsewhere that for some editors it's not enough to provide just the tool; Wikipedians also need information and instruction on what makes a good alt text description. If the proposal were to increase its focus on this aspect through the hub/project page and developing resources, I'd be more inclined to fund it.
  • Interesting problem to be solved, but my concern is connected with the tools. How will they be integrated? My opposition is mostly concerned with the clarification of the budget, as it is currently unclear.
  • I like the idea, but I want to make sure users like those mentioned on the talk page are consulted. I am also disappointed that some suggestions from the feedback last year weren't taken into consideration.
  • The project is interesting but the budget breakdown is non-existent. This is very crucial to understand: the hourly rates, or the breakdown of timelines. This tool should be implemented on Wikimedia servers for sustainability, given the scale of the workload.

-- MJue (WMF) (talk) 00:41, 3 June 2016 (UTC) on behalf of Marti (WMF)[reply]

Round 1 2016 decision[edit]

Congratulations! Your proposal has been selected for an Individual Engagement Grant.

The committee has recommended this proposal and WMF has approved funding for the full amount of your request, $14,000

Comments regarding this decision:
The committee is pleased to support build-out of a tool that will bring greater facility to the improvement of text alternatives to images. We heartily appreciate the focus on increased access for the visually impaired. The committee especially looks forward to seeing volunteers engaging with the tools created by this project.

Next steps:

  1. You will be contacted to sign a grant agreement and setup a monthly check-in schedule.
  2. Review the information for grantees.
  3. Use the new buttons on your original proposal to create your project pages.
  4. Start work on your project!
Questions? Contact us.

Good work[edit]

I finally located this grant proposal page and am really excited about what you are doing. I look forward to following its progress! Let me know how I can help support this or be a user tester if you need any! Hexatekin (talk) 22:54, 28 August 2017 (UTC)[reply]