Silver Community Group Teleconference

09 Nov 2018


jeanne, Charles, AngelaAccessForAll, Lauriat, KimD, Cyborg
jeanne, Shawn
Mike, jeanne


<jeanne> Minutes from AGWG - EU & Asia meeting where we did a presentation on Silver homework/Alpha Testing https://lists.w3.org/Archives/Public/w3c-wai-gl/2018OctDec/0062.html

Alpha Test as a term?

<jeanne> jeanne: People were asking if how "baked" the Silver prototypes are. What about Alpha Testing as a term?

<jeanne> Charles: It is ambiguous whether the testing is in Alpha, or the prototype is in Alpha.

<jeanne> Shawn: I think prototype testing is the most accurate, but it needs to be universally understood, which is isn't.

<jeanne> Charles: We are validating assumptions, so the next release is more likely to be the Alpha version.

<Mike> scribe: Mike

questions for feedback in the Silver prototype templates

Lot of presentations on the silver homework that we had given the AG Working group, talked about this at tpac but have not went through this in detail...until now! Have had some responses

jeanne: 2 new folders with information. 1 is in the simple language folder called INDIVIDUAL GUIDANCE DRAFT

<jeanne> https://drive.google.com/drive/folders/1yvZHLoAaXyK-MZ3CCYRBb75qPXiTowyj Folder of examples

jeanne: Another one for Information Architecture, had really good things here

<jeanne> IA folder https://drive.google.com/drive/folders/13ZUtgIDUyC8KTQcQqLzjqluDX7Z9fhmx

jeanne: Good article from Dave Swallow about the web of anxiety and accessible needs, link to a series of design patterns - have used one of these (Locate user in service journey)
... Met with low vision task force and used this as an opportunity to create example write up of "Point of Regard", created in plain language

<jeanne> https://drive.google.com/drive/folders/13ZUtgIDUyC8KTQcQqLzjqluDX7Z9fhmx

Mike: These are really good! Can start to populate prototype with real data now

<jeanne> Template for IA: https://docs.google.com/document/d/e/2PACX-1vRgf85Z_NJ7HmF-UX992wLx0F-sCQyipL6USL9HTmvBOWtH53C78SVNjKI8kLTxl5UuYJbc7ImiGsB_/pub

jeanne: Heres template for IA
... Need to look at questions and see if they are what we want them to be

Lauriat: Main point is not to answer these exact questions, its to write something in there for the template. We also want feedback on what the process was like. Think about process itself, what worked, what didnt, what they were happy with, what was hard

jeanne: I like those questions

Lauriat: Was it possible to create an outcome based on how you thought it would come out?

Charles: So did it match the template, was there soemthing that had to be added or removed? I'm missing the purpose of the template, this is for people participating in meetings that were had to write something within the template?

jeanne: Yes, these were all people that write SC and guidelines, they were user testing our prototypes

Lauriat: for the IA one we want to test the process of turning existing guidance or new guidance into new data that would fit into IA for silver and see how well that works

Charles: Is this mapped 1:1 on the IA prototype? Theres one for all the short description etc.

Lauriat: Yup, if someone was writing guidance that wasnt in silver yet this would be the data entry aspect of it. Right now in the prototype there is different sets of info because its in a DB and not flat file. The top level bit under guideline shortname has discrptions and related methods. The related methods is the join between descriptions and methods

Charles: Need questions on template itself. Looking at this as an outline one of the first things that I see is that there isnt an exceptions attribute under guidance or methods
... do people need that, expect that, is it useful?
... and the list of tags, is this appropriate of do we need to add / edit / remove

jeanne: We've been saying we have a seperate prototype of the tags

Lauriat: have been encouraging people to come up with their own if appropriate

Charles: Next Q, theres no definitions of what a method is? So can it be populated by what each of the things are

Lauriat: prefixed with a mini presentation at this point, so understanding is given before hand. E.g. under tests, we have to specify that these are the sorts of tests that would go in here
... not testing how the template works as data input, more how info moves from current guidance to this new shape
... need extra documentation for how it looks in its current form

Mike: What about things that should be covered?

jeanne: A lot is covered in plain language, but it could be asked

Mike: if it duplicated then we know that plain language is heading in the right direction

Responses for the prototype testing to date

Update on IA prototype

<jeanne> scribe: jeanne

MikeCrabb: I haven't had time to update more of the Information architecture prototype.
... I was at the @@ Minifest. I was showing the Information ARchitecture prototype at the conferenece for about 300 people. There was very positive feedback.
... these are people that have to comply, but don't like to because the information is so hard to find. They liked the tagging engine.
... Next step is putting some of the data that people we are creating into the prototype.
... I can move it to the W3C repo. The API is running on a simple PHP system that won't support much bandwidth.

jeanne: Is there anything we can do to move it to a faster system?
... I can approach W3C to put it on their servers? They will inspect it for security.

Mike: I can clean up the code, and I can also look at putting it on the university server.
... it can probably handle about 150 users at once.

Update on Plain Language

<Cyborg> just joined, sorry

<Mike> Angela: Wrote back to John a few days ago but thats the only activity to report on

Tim Boland's proposal on Conformance


<Mike> jeanne: Tim Bowland did a proposal for conformance complete with formulae

<Mike> ... thrilled that he has joined. He things really deeply about conformance with a great deal of expertise

<Mike> ... delighted that he things we should increase flexibility

<Mike> Charles: His proposal is similar to a heuristic evaluation, evaluator has to justify. So we need to have a justification process - whole thing hangs on this

<Mike> ... Burden is on tester to say why weights would change

<Mike> Mike: Very easy to game the system?

<Mike> jeanne: Could we do weights? One thing that people always want to know is why weights are set in a different way

<Mike> jeanne: How does weighting come in, do we say tests are weighted in a particular way?

<Mike> Charles: Not sure how we would arrive at a number but it would have to be a fixed value or range of value, dependant on number of tests

<Mike> ... I'm evaluating a web page, there are 10 possible tests that I can use, and 5 of them dont even apply. So does that mean that I can only possibly score a 10 out of 5 tests vs a 10 out f 10 tests? Do I have to adjust the weight or is the weight fixed?

<Mike> ... the weight of each test would increase

<Mike> ...A lot has to be defined, not just waiting on a potential score but human evaluation process..

<Mike> Lauriat: Another thing is gaming by adding things in 1 category by getting around certain tests which do apply. This is just looking at overall sum with no categorisation

<Mike> jeanne: We can add categories, we've been working on that. Know that we need some sort of minimum.

<Mike> ...this might be a much easy way to do the point system, this could be one point structure and people could do what they need to for that project within the categories

<Mike> Charles: Doesnt explicity say that score is achieved from human evaluator, just the rational. So automation testing and human evaluation is possible. Weight is fixed, and scores can be done by automation system / humans

<Mike> ...overall here is the result and here is why I used these tests

<Mike> jeanne: Some basic work has been done here, we can pursue this. Should we dig into it more?

<Mike> +1

<Cyborg> i can join back in if we go back into that process...

<Mike> Charles: Worth answering questions to take it to the next level

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.154 (CVS log)
$Date: 2018/11/09 20:02:23 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.154  of Date: 2018/09/25 16:35:56  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00)

Present: jeanne Charles AngelaAccessForAll Lauriat KimD Cyborg
Regrets: Jennison
Found Scribe: Mike
Inferring ScribeNick: Mike
Found Scribe: jeanne
Inferring ScribeNick: jeanne
Scribes: Mike, jeanne
ScribeNicks: Mike, jeanne
Found Date: 09 Nov 2018
People with action items: 

WARNING: Possible internal error: join/leave lines remaining: 
        <Mike> ... thrilled that he has joined. He things really deeply about conformance with a great deal of expertise

WARNING: Possible internal error: join/leave lines remaining: 
        <Mike> ... thrilled that he has joined. He things really deeply about conformance with a great deal of expertise

WARNING: IRC log location not specified!  (You can ignore this 
warning if you do not want the generated minutes to contain 
a link to the original IRC log.)

[End of scribe.perl diagnostic output]