- From: Nick Levinson <nick_levinson@yahoo.com>
- Date: Sat, 17 Jun 2006 18:25:17 +0000
- To: www-amaya@w3.org
If you ever update 8.8.51 for Win98 and/or if these are relevant to your later-platform version, here are critiques: 1. I look forward to complete compliance, although I assume that's difficult. I'm trying to find full support for ACSS or SSML so I can test my new pages, and so other browser makers will be inspired to follow in your footsteps. But I know that's a lot to ask. 2. During startup, the status bar says "Finished!". The time to say that is when it has executed everything, including loading the default home page, and is ready for the next input. Likewise when "Finished!" is displayed at times other than startup. Otherwise, I think I can do the next step, when I can't. 3. A progress bar for page loading would be useful. It took about 10-14 hours to partly load a 100K page of XHTML with no scripts and only minutes of multitasking but lacking alt attributes for many <img> tags (the text rippled every few minutes and shorter pages took a lot less to load). Loading was necessarily by forcing a character coding, ISO-8859-1, the same coding declared inside the file. Loading may or may not have finished; the screen saver was still not on after 14 hours. I ended the task, restarted Amaya, reopened the document, and noted that after it loaded the text above the fold (quickly above the fold) it then wrote each <img> tag (probably all defective in lacking alt attributes because the images are essentially invisible to users, being paragraph spacers). Each of these <img> tags took roughly 5-6 minutes to write (it paints <img> tags in 2 passes but the first pass may be quick before it unpaints them and paints them again). There should be 138 of the tags, which comes to roughly 11 1/2 hours to closer to 14 hours. I don't know if anything else has to be written or calculated after the <img> tags that would add notable time. The machine has adequate RAM, 64MB, unless Amaya has special requirements I missed, so a time estimate would be helpful for a long file. 4. The list of errors is incomplete. Fixing one error may lead a new one to appear, not caused by the fix. Apparently, the culprit is "mismatched tag". In one instance, it reported </div> as an error (probably as mismatched tag). That's fine, and I added the opening tag and saved (ignoring earlier errors); but then I closed the source code and error windows and reloaded the page (forcing 8859-1) and it reported a bad token several lines after the </div>, a token it hadn't mentioned the first time. If the W3C philosophy is to stop reporting errors until some others are fixed, this impedes page design if we opt to code a spot a certain way or can't identify the bad coding suggested by the errors list but want to know of other errors below it. I understand that tags that are not understood by a user agent are to be ignored, and meseems that an extra opening or closing tag should simply be ignored. 5. The "mismatched tag" error may refer to an error in tagging nested within. I had text within an invisible table cell, thus I had "<td> . . . <p> . . . . </p> . .. . </td>, except that I was missing the closing </p>. However, Ayama instead reported </td> as mismatched. Likewise, when Amaya reported </span> as a mismatched tag, I was actually lacking </abbr> within <span> . . .. <abbr> . . . </span>. Perhaps there's a better way for Ayama to report the inner tagging as erroneous. 6. I've also seen a "mismatched tag" error at </td> when I couldn't find an error. Within that table cell, I had a large volume of text, in one case as a single <p> . . . </p> enclosing several <br /> tags. Likewise, a </p> was reported as mismatched, but I couldn't identify what was allegedly missing, all the enclosed tags being matched, apart from one <br />, which isn't supposed to be. 7. In one instance, I suspect "mismatched tag" referred to a nonstandard attribute in the opening tag. The attribute, profile="[value]", is given by the Dublin Core organization's website and is useful, so I'll continue using it. 8. I've crashed it a few times via alt-space-x, which should maximize the window via the menu at the left end of the title bar, and once with just alt-space (I was going to test another command by keyboarding but didn't get a chance before the crash). That menu works fine via the cursor, the maximize button works, and alt-space-x works on Notepad. The crash symptom is the standard warning about the program having committed an illegal act and being shut down. 9. It crashes when (a) the main window was moved (so the maximize button could be reached) and maximized by the maximize button, (b) a file is being displayed, (c) its source code is open and maximized by the maximize button or is closed after being open, (d) its parsing error list is open or closed as long as there was a parsing error list and it was open, (e) I go to the main window, and (f) I select another file and click the dialog button to open it. (The file likely would have had to be opened by forcing character coding ISO-8859-1; that step would've been next.) The window turns unavailable, alt-tab does nothing, and clicking on menus or the close button yields only a click sound. The end-task dialog lists it but does not claim that it is not responding. I end the task and start Amaya again. However, if I had closed both the source code and the parsing errors list before opening a new file, I can proceed. Recapping regarding the child windows, either one or both being open leads to the problem, but if both are closed before trying to open another file the problem does not follow. On the other hand, if I edit the source code and save it, closing the child windows is not necessary to seeing the revised object-code page. 10. It reports "&" in a name attribute value as an invalid token and not well-formed, but not consistently. In one value, it reported the second "&" but not the first (they were nonconsecutive); in another, it skipped the "&" altogether but reported a similar error at the end of the name value, where I could not identify a nonalphabetic character or a wrong kind of quotation mark. 11. When forcing a character set is required, I'd like to know which character/s offended Amaya or if my UTF-8 declaration was misworded or inconsistent with file content (Amaya refuses an attempt to force as UTF-8 and prefers that I force as ISO-8859-1). 12. The Yahoo email login page (http://mail.yahoo.com) is mangled in Amaya. Maybe their page is nonstandard, and Yahoo doesn't support Amaya, but the page was so mangled I switched browsers. I'll use Amaya for page trials, since it looks good for that purpose. 13. I assume page rendering is not meant to implement all the coding, in order for Amaya to serve its purpose as a test bed for other aspects. Type size from CSS, for example, is not rendered. If that's intentional, it's not a priority for me, since other browsers can be used to test that aspect. 14. A minor convenience would be an alt-character choice for Force a Character Coding. I'm running it on Win98SE on a Compaq Armada 7800 model 6300/T/8000/V/0/1 with 64MB RAM and the resolution at 1024x768. -- Nick __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
Received on Monday, 19 June 2006 09:18:47 UTC