Doug, Thanks for your response. Since my post was critical of Window-Eyes, I'm copying this message to the same lists so others can consider your points. I mean to be objective, so I'll do some more comparing and report the results if significantly different. Regards, Jamal On Fri, 29 Jan 1999, Doug Geoffray wrote: > Date: Fri, 29 Jan 1999 07:36:16 -0500 > From: Doug Geoffray <[EMAIL PROTECTED]> > To: gw-info <[EMAIL PROTECTED]> > Cc: [EMAIL PROTECTED] > Subject: Re: A comparison of page loading times with IE5, Bridge, JFW, and WE > > Jamal, > > I appreciate your concerns but let me address a few issues which I feel are > very important. > > Before I get into the details I tried this same page myself. I am using a > Pentium II/333 with 128 megs of memory with a direct connection to the > Internet via a DSL line which gives me 384K down and 128K up. I am also > using the latest private build of IE 5. Beta 2 does have some major > problems with certain pages but these have been resolved with the build I > am using. Of course I don't have the other two screen readers so I can't > compare to them but I can certainly give you Window-Eyes information. > > I used the same page you were using http://www.samizdat.com. I used a stop > watch to compute the times so as you said, this is not scientific but > relatively accurate. > > The page took 7 seconds to physically load into Internet Explorer. I had > pictures turned off. Obviously it has to load before anybody can do > anything on it including the other two screen readers. I'm assuming you > did not take this value into account. The slower your connection to the > Internet the slower this value would be. My connection is very fast and > with graphics off, 7 seconds is about right. > > After the 7 seconds to physically load the page, the Window-Eyes MSAA mode > kicked in. It loaded in 26 seconds. If I turn MSAA mode off for whatever > reason and turn it back on for this same page, it loads in 18 seconds. > > This page contains 9,602 lines of information with 905 links. This is a > huge page. > > Okay, lets look at all of this. > > I agree my machine is much faster than yours. I also know you are going to > say not everyone has or should be expected to get my machine. But this is > the future of computers. What speed of machine did you have a year ago? > Also, 32MB of RAM for Windows 9x is a minimum with or without a screen > reader. Internet Explorer 5 is a memory hog. I personally believe > anything less than 64MB is a mistake with Windows 9x. Again, with or > without a screen reader. > > The latest build of IE 5 has certainly helped in the speed. Even with your > p/200 I wouldn't expect it to take 8 minutes unless you were counting the > physical load time and you had a bad connection and you had graphics turned > on. I would expect your machine to be around one minute. So maybe the > latest build makes that much difference or you had a very bad connection or > had other applications in the background eating processor time. > > I don't know your speed to the Internet but did you take into account how > long it took to physically load the page? Our MSAA mode won't kick in > until the page is completely loaded. So when did your timer start? Also, > when you pressed the JFW hotkey to reformat the page, was it already > loaded? In other words, did you include the load time with Window-Eyes but > not with JFW? Again, the load time could get large especially if you had > graphics turn on. I was under the impression you had to wait for the page > to load with JFW before you could hit the hotkey to reformat it. This is > why I'm thinking you didn't include the load time with JFW. > > The page you picked out put the most strain on Window-Eyes when compared to > JFW. I could pick a typical page with several frames that would put the > most strain on JFW. Again, with JFW you must manually reformat each frame > of the page. The page you selected didn't have any frames. But I'm sure > you will agree, more and more pages are using frames and I hope this > continues. Also, JFW has big problems with many sites that have frames. > They are just not readable. > > I don't know Window Bridge so I can only guess but I'm sure they are not > loading the entire page. So can you hit a hotkey to pop up all the links > like you can with Window-Eyes? Can you quickly go to the bottom of the > page? Can you do a find for something on the page at the bottom. My guess > would be no to all these questions. > > The bottom line, you can't pick one page and compare three screen readers. > You need to average all web pages together and not just pick a huge one > with no frames. Average small ones, medium ones, large ones, ones with > frames, etc.. > > But even this doesn't give the full picture. As I have told you in the > past, I personally don't believe any screen reader offers better reading of > Web pages than Window-Eyes. You mentioned this very briefly in your > message. But I believe this is a huge issue. Window-Eyes gives great > access to reading the web page. Now you said, "the Internet isn't just a > reading medium any more." And may be true, but there is a lot of it and > that will never change. > > In conclusion, I hear what you are saying. We are working with Microsoft > to decrease the speed of this load. But this isn't going to happen this > weekend. It will take some time but I think you will agree we are on the > right track. Window-Eyes is very fast with IE 3. Unfortunately, things > got very slow with IE 4 but things get much faster with each release of IE > 5. It still isn't as fast as IE 3 but IE 5 is doing a lot more work. > We'll get this down. Also, be careful making judgments with a beta copy of > IE! Things are changing all the time and you may not realize it. > > Doug > > > > > At 06:58 PM 1/28/99 -0500, you wrote: > >This is to report a simple comparison I just did of the time it takes to > >load a web page with Internet Explorer 5 beta 2 and Slimware Window Bridge > >2.54, JAWS for Windows, 3.2031, or Window-Eyes 3.0. My office computer > >(at a federal agency in Washington) is a Pentium 200 with 32 megs of > >memory and a direct connection to the Internet. > > > >I picked the home page of Richard Seltzer (http://www.samizdat.com), the > >author of Internet-on-a-Disk newsletter and various other publications. > >It is admittedly a large page, but thus also one that facilitates > >comparisons not skewed by small differences in network traffic and other > >variables. > > > >I first ran JFW and reformatted the page for optimum speech, resulting in > >a loading time of under a minute. I then ran Window-Eyes, resulting in a > >loading time of about 8 minutes. Finally, I ran Window Bridge, resulting > >in a loading time of under a minute. > > > >In short, the JFW and Bridge times were comparable and reasonable. The > >Window-Eyes time, unfortunately, was about 8 times slower and unreasonably > >long to wait, in my opinion. > > > >JFW does not use MSAA; Bridge and Window-Eyes do. The screen readers > >differ in the convenience of navigation and form entry once a page is > >loaded. This comparison did not address those other issues. > > > >Regards, > >Jamal > > > > > > Doug Geoffray > GW Micro, Inc. > Voice 219-489-3671 > Fax 219-489-2608 BBS 219-489-5281 > [EMAIL PROTECTED] or [EMAIL PROTECTED] > - Visit the jfw ml web page: http://yoyo.cc.monash.edu.au/~nallan/jfw
