SiteMorse 'report' (?)
Home / News & Resources / SiteMorse 'report' (?)
| Jon R wrote: |
| The 'filename' test was originally included because some HTML editors provide the filename of the image as the default 'alt' text - and that's obviously pretty useless as an equivalent. |
Useless according to the published guidelines, or in your humble opinion? As has been adequately illustrated there are circumstances when the inclusion of a filename in alt text is useful. Logical conclusion? The SiteMorse tests for alt text are flawed, and can produce false positives.
How common these false positives occur is irrelevant, what matters is that they are possible and you are aware of it, yet persist in publishing league tables based upon potentially flawed data. Really you should either fix or refine the tool or stop publishing.
| Quote: |
| It's already available - simply log on to the web site and look at the various report sections to see what is tested. |
Not what is being tested but how the testing is done, and what parameters and methodologies are used. Information like 'we fail alt text if it contains an image filename' would be ideal.
_________________
Dan Champion, Champion IS, Mooch Marketing, Revish
Share: 
Isofarro, you are not even pretending to be honest in this debate any more. Enough is enough. You are deliberately misreading and misrepresenting what I am saying and ignoring the substantive points so that you can ask the same question again in the pretence that I have not already answered it. There is no point in continuing this discussion.
Share:
| danchamp wrote: | ||
Useless according to the published guidelines, or in your humble opinion? |
Useless according to the published guidelines. They say that the alt text should be an equivalent to the image, not its filename.
| danchamp wrote: |
| As has been adequately illustrated there are circumstances when the inclusion of a filename in alt text is useful. |
Nobody has demonstrated that. As I said, in practice no circumstances, to my knowledge, have ever arisen where the filename is a genuine equivalent to the image.
Are you particularly expert at making mountains out of molehills or are you just excelling at it by chance today?
Share:
| Jon R wrote: |
| Useless according to the published guidelines. They say that the alt text should be an equivalent to the image, not its filename. |
No they don't, the guidelines make no mention of filenames. Out of interest why don't you test for alt text which is exactly the same as the image filename, rather than alt text which contains any image filename, anywhere within it?
| Jon R wrote: |
| Are you particularly expert at making mountains out of molehills or are you just excelling at it by chance today? |
It was my job, I used to be a lawyer, some scars never heal.
But I think you're doing the good folk here and myself a disservice by suggesting these issues are unimportant. They are not. Your practice of publishing 'league tables' based on your test results is likely to result in some people concluding that a site with a good SiteMorse rating is a high quality, accessible site. That would be a false conclusion.
Your colleague 'SiteMorse Access' admitted that your product can provide about 40% coverage for A and AA checkpoints. Even allowing him/her to round 32% up to 40%, that leaves a whole lot on the dark-side of your tests, doesn't it, and who knows what might lurk there?
I find your tool useful for validating my site, but view the practice of publishing league tables without sufficient warnings as to their true value a gross misrepresentation.
And FWIW IMHO Isofarro is right about the GAWDs slur - your company clearly and incorrectly listed them as failing both A and AA compliance. So far you have pleaded perfection - a little humility would go a long way.
_________________
Dan Champion, Champion IS, Mooch Marketing, Revish
Share:
| Jon R wrote: |
| As already discussed in this thread, I believe that SiteMorse was perfectly justified in flagging the alt text in question as problematic. |
Sorry Jon, but that statment is incompatible with a league table that is created where a factor that needs 'human assessment' is not assessed but classed as a fail.
Could the tool not be reconfigured to check for:
- existence of the alt attribute
- if alt text is solely the image reference, ie master.png, class it as a fail (I personally would prefer images to have better description than the file name, as I suspect do many others)
- if the alt text includes the file name and other text it gets excluded from the auto pass/fail and gets moved to the manual check
did I miss anything ?
Share:
| Jon R wrote: | ||
I don't have a copy of the original press release to hand, but it looks like you're taking a set of generic comments about a group of organisations, and for some reason reading it as if they're all directed personally against GAWDS - when it's fairly clear that's not the case. |
No Problem, you can get it where I got it: http://media-insert.co.uk/..., this has not yet been adjusted as the text at SiteMorse has been. Not have I suggested it was pointed at GAWDS, these comments are aimed at the core sites of accessibility, one can only but understand these comments to be pointed at the list of sites that follow. GAWDS is at the top of that list. So yes I take it these comments were targetted at GAWDS as much as at the other groups like WebCredible etc.
| Jon R wrote: | ||
Because, false positive or not (and as I already mentioned, I don't agree that it was), it's only one minor point and not worth having a big argument over. |
But it was the one minor poit that landed GAWDS on the list of sites that are not accessible, so hence it is hardley minor.[/url]
_________________
Kyle J. Lamson
Analyst/Programmer III, State of Alaska
--
LSW-WebDesign.com & DarkShadow-Designs.com
Share:
| Jon R wrote: |
| Isofarro, you are not even pretending to be honest in this debate any more. Enough is enough. You are deliberately misreading and misrepresenting what I am saying and ignoring the substantive points so that you can ask the same question again in the pretence that I have not already answered it. There is no point in continuing this discussion. |
Resorting to ad hominem attacks because you disagree with someone else's viewpoint is not really acceptable - either address the perfectly valid points raised and prove he is wrong, or admit that your tool has an error.
_________________
The Watchmaker Project - my personal blog
29digital Design Studio - freelance web design/development
Share:
| danchamp wrote: | ||
No they don't, the guidelines make no mention of filenames. |
Your argument would appear to be that its filename has a noticeable probability of being an equivalent to the image. My argument is that that is highly unlikely indeed, and in practice never, or almost never, happens.
| danchamp wrote: |
| Out of interest why don't you test for alt text which is exactly the same as the image filename, rather than alt text which contains any image filename, anywhere within it? |
Partly because the image may have been renamed since it was put on the site, and partly because that has never been necessary because, in practice, false positives do not seem to occur due to this test.
| danchamp wrote: |
| Your practice of publishing 'league tables' based on your test results is likely to result in some people concluding that a site with a good SiteMorse rating is a high quality, accessible site. That would be a false conclusion. |
I am not sure how many times we have to say that automated testing, although a useful tool to help you create and maintain accessible web sites, is not capable of getting you all the way there by itself, before people stop accusing us of trying to claim otherwise. It is not, and has never been our position that automated testing is sufficient by itself.
| danchamp wrote: |
| I find your tool useful for validating my site, but view the practice of publishing league tables without sufficient warnings as to their true value a gross misrepresentation. |
I am not sure what you would consider "sufficient warning". As far as I am aware all the league tables including a note to the effect that automated tested does not cover all or even most of the guidelines. A very important point that people seem to be missing also is that it is not an accessibility league table! Accessibility (as measured by automated testing, of course) is only one of the factors that is taken into consideration when ranking the sites.
| danchamp wrote: |
| And FWIW IMHO Isofarro is right about the GAWDs slur - your company clearly and incorrectly listed them as failing both A and AA compliance. |
I don't agree that it was incorrect - but this is definitely where we get to mountains and molehills in my opinion. There is only so much debate it is worth having over the alt text on a single image.
| danchamp wrote: |
| So far you have pleaded perfection - a little humility would go a long way. |
Since I have specifically claimed imperfection, I am not sure how you can possibly say that!
Share:
| elfin wrote: |
| Could the tool not be reconfigured to check for:
- existence of the alt attribute - if alt text is solely the image reference, ie master.png, class it as a fail (I personally would prefer images to have better description than the file name, as I suspect do many others) - if the alt text includes the file name and other text it gets excluded from the auto pass/fail and gets moved to the manual check did I miss anything ? |
Yes, that is of course possible. I could probably arrange that - but it would help if someone can come up with an existing real-world example where a filename plus other text is a genuine equivalent to an image in a web page. I am perfectly willing to be persuaded on this, but I need a rational argument based on reference to the WCAG 1.0.
Share:
| Jon R wrote: |
| Isofarro, I am not interested in picking pointless nits with you. Several people expressed an opinion that it would be good if we engaged in a constructive dialogue in this forum. This is not a constructive dialogue.
If people want to discuss the SiteMorse product and ways in which they feel it could do better or be improved then I am happy to do that - and I am in a position such that I can probably do things about it. PR and marketing are not my department and I have no particular say over them. Why you have to try and defend the GAWDS website as if it must be absolutely perfect I don't know - perfection is extremely rare! If the particular alt text we have been discussing is the closest thing it has to an accessibility problem then it is doing well indeed. |
Ok, that I can deal with. I am a proud member of GAWDS and have seen the first sites quoting this report as fact to show GAWDS and such organiations are not thir worth.
I agree we have to issues, your product where you can speak, your companies business tactics which you are trying to seperate yourself from. Understandable as yu would seem to be developement team.
However you can thank your PR people for binding the two areas together. By claiming GAWDS fails A & AA tests, they claim your product is trustworthy. One cannot seperate them anymore as GAWDS is a example of the problem with automated testing.
We claim your tool has a problem, you feel it does not. We must show examples of why their is a problem so we must discuss the GAWDS Rating. But this is not a private discussion, GAWS was never told it was being tested as claimed by SiteMorse, so we could not discuss it normally. Your PR section/Marketing has jumped the gun (if tested sites are trully notified) by pubicly stateing GAWDS and the others have failed A & AA testing based on what we now feel is a error in your tool. GAWDS has a reasonable alt even without the image name, so the image name does not change that, at best it cabe considered "extra info".
As much as I understand your not wanting to venture into the marketting goof up, it cannot be avoided as that got this whole thing rolling.
If SiteMorse Access is Marketting, and you are developement and we already have contact with Mr. Shaw, maybe a actualmeeting can be organized with representatives of GAWDS and other organizations listed and you and representatives of the other SiteMorse sections to discuss this. This way it can be cleared up and you may get the technical suggestions you wish for without being in the position to have to defend what Marketing did wrong.
Perhaps you can suggest this to your bosses and have a round table meeting. There are also many who feel this whole legue table business is not good as managers require developers to tweak the site to rate well even when not accessible. I am sure this SiteMorse "town meeting" with those who make up the accessible community would be welcomed from all sides.
_________________
Kyle J. Lamson
Analyst/Programmer III, State of Alaska
--
LSW-WebDesign.com & DarkShadow-Designs.com
Share:
| Jon R wrote: |
| Your argument would appear to be that its filename has a noticeable probability of being an equivalent to the image. My argument is that that is highly unlikely indeed, and in practice never, or almost never, happens. |
No, my argument is that the guidelines don't discount the use of an image filename in alt text. The burden of proof here is on you, and you alone as the publisher of this information. It's not for us to provide examples of when an image filename might legitimately appear in alt text, but for you to prove that it will never legitimately happen, since that's the interpretation you've chosen to place on the guidelines.
To paraphrase a most excellent movie, "I don't believe in god but I'm afraid of him". You seem to be close-minded enough to believe that because you haven't ever seen it, it has never happened in the past and will never happen in the future. That's quite a risk, and a very worrying approach for SiteMorse or any software developer to take.
| Jon R wrote: |
| in practice, false positives do not seem to occur due to this test. |
What, you have personally checked the gazillion errors SiteMorse claims to have found on the sites it has tested, and none of them was a false positive for this checkpoint? That's an impressive claim.
| Jon R wrote: |
| I am not sure how many times we have to say that automated testing, although a useful tool to help you create and maintain accessible web sites, is not capable of getting you all the way there by itself, before people stop accusing us of trying to claim otherwise. It is not, and has never been our position that automated testing is sufficient by itself. |
Well, if you have to keep saying it maybe that's because you're not saying the right thing? Perhaps you should be saying how little of the way there it gets you.
| Jon R wrote: |
| I am not sure what you would consider "sufficient warning". As far as I am aware all the league tables including a note to the effect that automated tested does not cover all or even most of the guidelines. A very important point that people seem to be missing also is that it is not an accessibility league table! Accessibility (as measured by automated testing, of course) is only one of the factors that is taken into consideration when ranking the sites. |
As others have noticed you seem not to be reading some of the important words on this thread. I clearly referred to 'high quality, accessible websites' for the very reason you mention. It's irrelevant what weighting you give to accessibility, if your methods and conclusions are flawed so are the composite results. Sufficient warning for me would be a statement declaring the margin of error in your accessibility tests - 120% or thereabouts. Blimey, when you write it like that it's a bit scary, isn't it? Even the election pollsters manage a margin of error of just 3%.
| Jon R wrote: |
| There is only so much debate it is worth having over the alt text on a single image. |
I've derived great value from the debate. It has informed me no end about SiteMorse, its attitude, approach and values as a company, and what it believes is important. Were we just debating the alt text on an image? I've come to treat these threads as a proxy indicator for SiteMorse.
I'm checking out of this discussion now. Thanks for participating though Jon, it's been enlightening.
_________________
Dan Champion, Champion IS, Mooch Marketing, Revish
Share:
| Jon R wrote: | ||
Yes, that is of course possible. I could probably arrange that - but it would help if someone can come up with an existing real-world example where a filename plus other text is a genuine equivalent to an image in a web page. I am perfectly willing to be persuaded on this, but I need a rational argument based on reference to the WCAG 1.0. |
Currently my other half has been forced into a position where she has to use Voice Recognition Software. One of the things she has noticed is that long link text are a problem to use. That is one example.
Reading http://www.w3.org/... Guideline 1
| Quote: |
| Unless verbal descriptions of this visual information are provided, people who cannot see (or look at) the visual content will not be able to perceive it. |
an image by its very nature is graphical and conveys a lot of information. Repeating all of it, visibly, on a page might not always be practical. By adding in the filename to a link to download an image you aid anyone that wishes to download the said image. The extra information provided within the alt text, especially in the example we are all on about, may actually be advantageous, as it provides information that 'sighted' users can see but 'non-sighted' users cannot. By its nature it may also provide an acceptable alternative for those using non graphical browsers.
(I may be mistaken on a point or so in that but the gist of it remains valid)
Share:
| Jon R wrote: | ||||
Useless according to the published guidelines. They say that the alt text should be an equivalent to the image, not its filename.? |
There is nothing in the guidelines, as far as I can see, that specifically bans the use of a filename if it could be considered part and parcel of the content information.
The WCAG Core Techniques document further elaborates on alternative text as follows:
| Quote: |
| When a text equivalent is presented to the user, it fulfills essentially the same function (to the extent possible) as the original content. For simple content, a text equivalent may need only describe the function or purpose of content. For complex content (charts, graphs, etc.), the text equivalent may be longer and include descriptive information. |
http://www.w3.org/...
This would imply that, if the inclusion of a filename is appropriate when describing the function or purpose of an image, it would be valid under WCAG 1.0.
I would agree that developers who use filenames as default alt text without either thought or consideration need to have the potential accessibility issues brought their attention. However , there is a big difference between that and ' failing' sites that have chosen WCAG compliant alt text with due care and attention.
| Jon R wrote: |
| Are you particularly expert at making mountains out of molehills or are you just excelling at it by chance today? |
In all fairness, the original mountain was of Sitemorse's making. I couldn't agree more that this issue is really something of a minor molehill. What does, or does not, constitute informative alt text likely to be a topic for perennial discussion. And there's nothing wrong with discussion. We can all learn something from each other. But using a molehill to try and publicly undermine someone else doesn't help anyone. And even worse, it creates negative publicity for the whole field of Web accessibility - which is something we could all do without.
_________________
Mel
If you notice rather bizarre typos, it's because I'm not typing but using Voice Recognition software. A spill checker doesn't do any good. Please have patients. I do attempt to edit my coffee but occasionally mist
Share:
| Jon R wrote: |
| Yes, that is of course possible. I could probably arrange that - but it would help if someone can come up with an existing real-world example where a filename plus other text is a genuine equivalent to an image in a web page. I am perfectly willing to be persuaded on this, but I need a rational argument based on reference to the WCAG 1.0. |
Well, I've come across three companies called "PNG" - all with web sites -1 website for the JPEG group and another website for a company called "BMP". It's not completely impossible for a company to have a web site logo/name that superficially resembles a file name. Under these circumstances, it would not only be considered appropriate to have alt text that resembled a filename but possibly mandatory , if it echoed the text displayed within the image.
Also, on deeper reading of the documents that accompany WCAG 1.0, it would seem to be important that the function of an image be considered when creating alt text and not, necessarily, a simple text replication of its content.
Consider a page containing a number of graphic images for download. When separated from their containing page (ie downloaded), the only thing that differentiates them, from the point of view of someone with a visual impairment, is their file names. Would it not make sense to provide the filenames as part of the alt text so that visually impaired users had a common point of reference between their downloaded image(s) and the appropriately captioned images on the web page?
Next Scenario:
A web page intended to illustrate the visual differences between different graphic file types. Sample images could be created that contained text such as "sample.gif" and "sample.jpg" etc. Best practice and compliant alt text would replicate the text within the graphics.
Finally, a screenshot of an application contains a few words of text that includes an image filename. Again, the accompanying,valid, alt text would contain a reference to a file name and the page would "fail" according to the current Sitemorse tool.
I wouldn't dream of claiming that these examples are commonplace. In fact, they would probably only occur once in a blue moon. But the fact that these examples could occur suggests that either the Sitemorse tool needs adjusting, or that its results may need to be manually verified before jumping to conclusions.
_________________
Mel
If you notice rather bizarre typos, it's because I'm not typing but using Voice Recognition software. A spill checker doesn't do any good. Please have patients. I do attempt to edit my coffee but occasionally mist
Share:
| Quote: |
| SiteMorse reveals that a number of providers, or experts in the field, fail both the mandatory accessibility standards of A and AA. |
Mandated by whom?
Share:
All times are GMT
You cannot post new topics in this forumYou cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum