[EPUB3 Grid WG] Global and Fixed Layout Tests

[Ric Wright]

I have just spent the last few days running through the tests on the Readium platforms (iOS, OSX, Chrome, Android and Cloud are pending).  Two areas caused me some difficulties, which I thought I would share with you.  

BTW, let me preface this by saying I am not clear on which mailing list comments/discussions like this should go, Epub3-Grid, or TestSuite (so I cc’ed both). Can someone indicate which I should address these issues to?

First, the Global Language files Japanese test files (EPUBTest 0130) are hard for someone like me who doesn’t speak or read Japanese (English, French and a little German is my limit).  Fortunately, some kind soul has added images (bitmaps).  I want to make clear that I thank those who put this together and thoughtfully provided the bitmaps.  Just trying to make it even more efficient.

However, one of the issues I ran into is that the bitmaps that show “how it should look” are formatted such that only if the RS window has the same proportions (sort of letter-sized landscape) do the (reflowed) results look (sort of) the same – I.e. It is hard to get the line-wrapping the same.

I haven’t taken the EPUB apart yet, but wouldn’t it have made sense to either 

• Have some verbiage in the prologue of the document specifying how to format one’s browser so the layout of the rendered form looks the same as the screen capture? Or,
• Wouldn’t make sense to format the EPUB as a fixed-layout so those of linguistically challenged types don’t have to try to twiddle their windows (which can’t even be done on a tablet) to get the wrapping of the lines the same?

Second, the fixed layout tests are rather hard to interpret in many cases since the only content on the pages is the page number at the bottom.  So whether a page is or should be portrait or landscape is often a little difficult to discern. Unfortunately, we have several bugs in this area which complicated my testing.  I ended up having to take the tests apart and add some content (fixed sized divs with a border, etc.) in order to try and figure out what was going on.

I would love to volunteer to help out with this but am already under water, but I will help as I can – assuming anyone thinks these are valid points.

Thanks

Ric

[Patrick Keating]

I think your points are extremely valid, Ric.  I am not sure what the difference in the two mailing lists are but I am glad to be on at least one of them: QA on Readium is our greatest priority now imo.

FXL is extremely hard to test and I know Kathy has been making some good progress on how to test it.  We should have this and the build system be the first thing that the new QA person looks at when they start.

Let us know if we can help with anything until then,

Patrick 

--
You received this message because you are subscribed to the Google Groups "epub-testsuite-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to epub-testsuite-di...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Patrick Keating

CTO, Bluefire Productions

Seattle, WA

(206) 781-1411

[Ori Idan]

First of all, the right mailing list is the TestSuite mailing list, since the IDPF test suite group is responsible for these tests.

I hope that one of the Japanese members who wrote the Japanese tests will comment on your mail.

Anyway, I will look into the tests myself in the next few days.

-- 

Ori Idan CEO Helicon Books

http://www.heliconbooks.com

[Ori Idan]

[Kathy Alley]

[matt.g...@bell.net]

Hi Ric,
 
It's probably more appropriate to send these sorts of problems to the test suite's mailing list, since that's where the developers of the content are. If the bug(s) will materially affect the reporting of results, it'd be good to include the grid group, of course, but doubt we can solve them there.
 
We've kind of been rushing bug patches to the tests to ensure the integrity of the grid, but when we get into improvements that require some expert knowledge and/or more work to implement they should go into the issue list so they don't get forgotten or missed by one or the other group (esp. since not everyone can cross-post or follows both).
 
I can't offer much help on the rtl/vertical-writing tests, unfortunately. We'll need to engage the original authors with your suggestions.
 
The FXL tests, as I'm sure you can tell, were still under development at the time we cut off work on them so we'd have a reasonably stable set of tests. It's definitely on the todo list to add real content, finish the prologue pages, etc. but we ran out of hours in the day.
 
Regards,
 
Matt
 

---

Date: Mon, 24 Mar 2014 11:24:46 -0500
Subject: [EPUB3 Grid WG] Global and Fixed Layout Tests
From: rkwr...@geofx.com
To: epub-testsu...@googlegroups.com; epub3-grid-w...@googlegroups.com

[Ric Wright]

-Grid WG

Matt,

Thanks for the feedback.  Let me know if I can be of assistance.

BTW, I have attached one of my modified FXL Test Suite files.  What I changed was:

• to each test page
• Added div  which contains
• A sentence stating what that particular page should do
• A bunch of paragraphs of lorem ipsum
• A div which has 4 divs to outline the page
• A link the CSS file (fixed.css)
• Added a CSS file fixed.css which contains
• A text container style (red outline, grey background)
• The CSS to style the 4 divs to outline the page
• A style to emphasize and center the sentence

These changes made my testing much easier (across our 6 platforms).

Ric

[Kanai, Takeshi]

Hi Ori and Ric,

 

I’m another Japanese member in TestSuite mailing list, and I’ll modify the file and make it easy to use.

 

 

In my view, the first trap which makes you nerves is pagination-rtl-030 and 040. Bingo?

What testers need to do in these tests are to see whether the order of both A, B, C in red and circled 1, 2, 3, 4, 5 are aligned with “Expected display image”. So, if you see “B” block after “A” block, and “1”, “2”, “3” inline block with the same order of the image, then the test passes. The place where you can see the circled numbers are not a matter of the evaluation, although that is not clearly mentioned.

 

 

One fundamental issue which makes the creation of Japanese test case difficult is that most of testers who are not familiar with vertical writing layout need to change the orientation of the device to read it. Then, top, bottom, right, left, direction in the test file is not the same with what the testers see. When it comes to PC/Mac viewer testers, they need to endure the neck pain.

 

To be honest, I can’t find any critical reasons to use Japanese text for several test cases and I would like to remove such language restrictions as much as possible.  Do you agree?

 

I’m afraid that it will take few weeks to modify the test cases. In the meantime, I would like to tell you the intention of each test case as much as possible. So please indicate which test cases are difficult to read.

 

 

Thanks,

Takeshi Kanai

[Ric Wright]

Kanai-san,

Thank you very much for taking the time to make these changes.  Much appreciated.  No rush.

Best regards,

Ric