Inside IE 8 RC1 with Dean Hachamovitch and Jason Upton

The Discussion

• 

  So I hope  Dean Hachamovitch or Jason Upton or anyone else in IE team reads this because I want to tell them, and I'm sorry for harsh words, that if I were their boss I'd fire them. Because you absolutely deserve it for not fixing the 'bmp pictures' bug.
  
  An hour ago I've downloaded IE8 RC1 on 32 bit Vista Business with SP1, fully updated and what bothers me is a bug that was partially responsible for me switching to an alternative browser.  When you right click on picture and select 'Save As Picture' the only available format is bmp file.
  This bug cannot be always reproduced, it's affected by header information that comes with picture, but personally I can always reproduce this bug on this site. I've looked for a solution on internet, but it seems they are all written for older versions of either IE or OS - deleting temporary files doesn't work and it's not possible to empty 'Downloaded Program Files' folder under Vista as is suggested in KB article.
  
  Now I've done some research (read googling) and it seems the error is caused because IE sometimes (I guess probably for perfomance reasons - to save memory in ram) discards original format and keeps only bitmap representation that is convenient beacause this is format that can be displayed with the least amout of cpu processing. I'm fine with that, but why couldn't you redownload image when user selects 'Save As Picture' menu  - or at least convert it back into original format if resending picture file is not possible. If you click 'Properties' on that same menu, IE reports original format (so IE certainly has some knowledge about format of original format), so why not convert it back - I don't really mind artifacts that will come with encoding back bitmap to original lossy format (like jpg).
   
  I can't stand BMP files on my disk - you may think it's good enough, it gives you few megabytes smaller memory footprint (but you already eat tens of megabytes), problem is everybody else does it correctly - Firefox, Opera, Chrome and others. 
  
  You could say I'm a bit of ms fanboy, you don't see me using that tone often - but really - it's either you are doing this intentionally and aiming for cheap way to lose few megabytes in memory with too aggreessive cache regime , or - and there is argument for this, you are not doing this on purpose and are just too incompetent to fix this bug that, according to previously mentioned KB article, has been plaguing IE since July 2000 - when IE 5.5 was released. 8 years.
  
  Again sorry for angry words, but it's about time you fix this bug - otherwise IE is great, I love accelerators.

• 

  Dean here -- Jason and I are sitting next to each other (again). We read your post. We're looking into the issue now. Thanks for pointing it out!

• 

  NERD RAGE OMG YAY!!!!!!!!
  

• 

  I'm not sure what to complain about: IE saving as bmp or that site sending images with a pragma: no-cache header. I think I'll complain about the second .
  

• 

  the bitmap bug (once again) verified on vista 64 ie8rc btw.. but that is really strange bug imo.. it seems like something that wold be so trivial to fix, the app already knows the url to the image O_o and even if we ignore that, how hard is it to convert something to a jpg? frickin paint can do that..
  
  im a ie fan i gotta say and im really happy the zoomed-in perf has improved in the rc what i dont get is why the ie team doesnt use .net more.. imo, it should all be managed code.
  also, memory consumption is just silly.. and avg of 50 (fifty!) megs for one tab..? come on.. sure thats the private bytes and taskmanager is not exactly a great profiler.. but still, that really taks a lot of explaining..
  
  ie has come a long way but i still feel microsoft could really do even better :/
  
  
  --edit post watch--
  this is a really interesting video imo, dont devalue it guys ie is a very central bit of software and there are a few things in the design of iw that seem strange to the outside observer.. because of that, i find it really interesting to hear how the ie team thinks about stuff it would be interesting to hear about the i.e teams oppinion of .net and why they haent adopted it more, im sure there is a reason (though i still think ie would be better of if it was .net )

• 

  Acid 2 was mentioned. What about the Acid 3 test? IC8 RC1 gets 12/100, Firefox 3.0.5 gets 71/100, even tried it in Flock 1.2.7 and it gets a 53/100.

• 

  Care to show me a browser that doesn't use 50 megabytes for a page (tab)? . It's funny because I keep seeing browser users complaining about their browser (IE, Firefox etc.) consuming a lot of memory but if you open the same web page in all of them and then take a closer look at various memory numbers you see that at the end of the day they all use tens of megabytes of memory. For example https://channel9.msdn.com (private bytes usage as reported by ProcessExplorer running on Vista):
  
  IE8RC1: ~37 mbytes
  Chrome: ~36 mbytes
  Firefox: ~54 mbytes
  Safari: ~50 mbytes

• 

  i dont think thats a valid reason for consuming that amount of memory so what if the current generation of browsers are all memory hogs.. that doesnt change the fact applications that have more complex ui require much less memory..
  
  --edit--
  
  this is atleast how it seems to me i just think its strange that a relativly small amount of html and js and other fluff can take up so much memory in such  a consistent fashion
  
   id love to hear the ie team talk about why im wrong
  
  --edit2--
  
  fyi i get ~45mb when going to channel9 i also noticed that the memory doesnt go down when i leave a page, but privte bytes dont tell the full story so maybe thats nothing to worry about..

• 

  Yay IE8 zoom in XP works without completely bogging down the rendering!  RC1 is way more polished than the beta. 
  

• 

  Keep the (constructive) feedback coming! The IE team is listening.
  
  C

• 

  It's a shame this won't be available for Win7. I hope an updated Win7 build with newer IE8 bits will be released soon to the testers, but I'm not holding my breath. It really is a pity as IE8 in Win7 is almost unusable, like beta 2 was. I'm hoping that RC1 fixes all the crashes, hangs and the plethora of rendering issues. I'm installing it in a virtual machine for some testing, but meanwhile I'm stuck with a terrible IE8 build in an otherwise excellent Win7.
  
  EDIT: At least my sidebar gadgets work again with RC1, and my site finally renders correctly (in the build in Win7 it's even broken in compatibility mode).

• 

  Thanks for nabbing the IE guys for an interview, Charles!

  By the way, here's some markup and CSS that doesn't work right in IE 8 RC1 in EmulateIE7 mode but works in IE 6, IE 7, IE 8 beta 2, and IE 8 RC1 in IE 8 strict mode.

  <!DOCTYPE html>
  <html>
  <head>
  <title>Centered Table</title>
  <meta http-equiv="X-UA-Compatible" content="IE=EmulateIE7" />
  <style type="text/css">
  body, table, td { margin: 0px; padding: 0px; }
  body { font-family: "Segoe UI", sans-serif; margin: 1em; }
  table { border-collapse: collapse; margin: 0px auto; }
  td { background: url('https://i2.microsoft.com/en/shared/templates/components/cspMscomMasterNavigation/m.m.nav.png') repeat-x; }
  td { border-right: dotted 1px #000; }
  td { padding: 0.5em; }
  td.nosep { border-right: none; }
  </style>
  </head>
  <body>
  <div>
  <table>
  <tbody>
  <tr>
  <td>Windows</td>
  <td>Office</td>
  <td>Partner & Customer Solutions</td>
  <td>Support</td>
  <td class="nosep">About Microsoft</td>
  </tr>
  </tbody>
  </table>
  </div>
  </body>
  </html>

• 

  Don't try installing this version if you're running Windows 7—you'll just get a message that says "Internet Explorer 8 is not supported on this operating system." This makes sense, because although IE8 will be the browser that comes with Windows 7, the RC1 installers are for Vista (in both 32- and 64-bit flavors) and XP. It does, however leaves us in the unusual position of running a newer version of the browser software on the older OSes.
  
  source http://www.extremetech.com/article2/0,2845,2339618,00.asp?kc=ETRSS02129TX1K0000532

• 

  The silliest comments I have seen in a long time... If a site says no-cache and IE don’t cache then fine.

• 

  Does this actually work?
  
  
  
  Posting from Leeds in Yorkshire by the way.
  

• 

  Something that always bugged me about browsers such as Firefox was that the rendering of certain elements with CSS would flicker and change with a mouse hover, even when the css had no mention of mouse events. Obviously this was due to the way they managed CSS internally. IE6 and 7 were always the exception to this, if something rendered and you hovered, that's the way it stayed. I was quite worried to see IE8 fall into to this strange habit (admittedly perhaps the CSS was poorly written, but it's not the point, it shouldn't flicker and change once rendered). Could you perhaps explain why IE8 works this way?

• 

  Could you post some examples of this flickering?

• 

  Hi,
  
  I've got this senario: IE 8RC1 running on Windows XP. In 1 tab I'm on a website (say, channel 9). In a second tab, I'm on Windows Update. When I click the 'Install Updates' in the second tab, a download box appears showing the progress of the download and installation. However, I can not navigate back to the first tab while this box is still there.
  
  Why can't I leave Windows Update running in one tab and continue my 'browsing experience' in another?

• 

  Upgraded to RC1 and now I'm getting prompted for the Adobe FlashPlayer where I previously was not.
  
  I have the flashplayer installed but disabled for all sites other than those I specifically enable it for (because I generally hate flash).
  
  Anyway, I reported it, along with a screenshot of the C9 homepage asking for Adobe Flash to be enabled.

• 

  That's because of a script many Microsoft sites use for Flash installation and version detection.

  var i,flash;
  if (window.ActiveXObject){
      for(i=10;i>0;i--){
          try{
              flash=new ActiveXObject("ShockwaveFlash.ShockwaveFlash."+i);
              WT.fi="Yes";
              return i+".0";
          }
          catch(e){ }
      }
  }
  

  
  

  The non-IE branch does this in a better way, querying the navigator's plugins collections to determine which version (if any) of Flash is installed. IE needs a better way for scripts to query if certain ActiveX objects are available. Otherwise, scripts like this will (incorrectly) trigger IE to report that a page uses Flash. The way this script conflates detection with instantiation is ugly.

  This is probably one of those "damned if you do, damned if you don't" scenarios. Exposing this collection, even a "safe for scripting and user-enabled only" subset, would just make it easier for those malware (expletive)s.

• 

  Add ons disabled?
  
  Surfing around the net I get the yellow bar "Internet Explorer is currently running with add-ons disabled. Click here to manage, disable, or remove your addons."
  
  So I click Manage. I see a bunch of them like Flash are enabled. So a bunch of addons are enabled. Am I missing something? Should there be a checkbox to enable addons or something?
  
  Running: Vista Home Premium on Motion LE1600 Tablet PC

• 

  Did anyone get flash working? Is this an Adobe problem that will be fixed?

• 

  I tried printing this article http://en.wikipedia.org/wiki/Proton_NMR to which the second page on print preview and during actual printing is blank. It works fine on Firefox and I believe on the previous version of IE.

• 

  I think you have done a good work on IE8 but I think it is not enough 'cause css engine has still some annoying bugs.
  One bug is: It's not possible to make a background image on a table via css (using background-image rule); another bug is: position: fixed is unsupported. I have also seen that site properties' page is legacy of IE5.x
  

• 

  position fixed works even in ie7.. and you can set background-image on a table fine, although I find that questionable use.
  
  edit: im more confident about ie8 now because it can actually render again vs the beta which had pretty lousy rendering.

• 

  Ah... no.
  
  While I can't vouch for what C9 is doing in it's page init scripts, I know for a fact (because I wrote them) that some of the sites prompting me for flash do not use or check for flash, or any ActiveX objects.
  
  Moreover, the point was that up to Beta2 everything was fine, only with RC1 am I being prompted to enable FlashPlayer for a whole bunch of sites than previously didn't require it.

• 

  i noticed that on some pages, scottGu's blog for example, the text doesnt re-layout properly when you change the text size (page > text size) if the page is erloaded or even if the browser window is resized, the text gets layed out again and looks alright.
  text in a paragraph also get re-layed out if i click on it
  more specifically, its the line hight that doesnt seem to get updated properly
  other than that i gotta say RC is very polished, nicly done again, zoomed in perf is great now
  
  got one question though, why doesnt the ie team leverage the dlr js engine? it kicks butt pref wise.. i'd love to hear a discussion on that

• 

  Are you sure it's about layout and not rendering? It looks to me (yes, I ran into the same issue) that's the same bug from the beta versions where parts of the page were not rendered until you clicked them. Kind of scary for this to happen in a "release candidate".
  

• 

  no, im not sure at all they just dont look right
  i do think it looks like the line hight doesnt get updated properly, because if you go from larget text to smaller, the lines are as far apart as they where with the larger text

• 

  I find the refusal to provide support for border-radius to be very frustrating.
  
  http://msdn.microsoft.com/en-us/library/cc351024(VS.85).aspx">http://msdn.microsoft.com/en-us/library/cc351024(VS.85).aspx
  
  Obviously this is CCS-3, but it is *SUCH* an awkward thing to manage otherwise, and the other major browsers now make it so much easier with their implementations of it... that I can't understand MS forcing devs to continue battling with this issue for how ever many more years before we see an IE9

• 

  I would also have liked to see a way to uninstall addons from IE... when those horrid things won't provide a standard remove option for you to run. Some crap like Adobe Acrobat addons are just totally recalcitrant, and won't give you the option to uninstall, and there is not option in IE to do so. Sure, you can disable them, and you can probably screw about in the registry editor and hope you get it right, but why can't there be some basic "get rid of this crap from my IE" options in IE ...similar to the way you can remove ActiveX controls from IE.  ??

• 

  Mine were not enabled until I had the latest version of live addons.  Make sure your addons are upto date, it seems addons will be disabled until they are.

• 

  This is a funny bug.  I didn't know of this until RoyalSchrubber mentioned it, but now that I know about it, I'm noticing it.

• 

  The configuration dialog is a real downer.  Trying to force people to use the settings Microsoft wants is a disgrace.  If I see the dialog again I may think about switching to Mozilla and I am a fairly hard core Microsoft supporter.

• 

  I was just commenting on the reason why it happens on Channel 9.  Since I have no idea what other sites you were talking about, I wasn't commenting on them.

• 

  Sorry, but you are still missing my point, which is that it's only since RC1 that this has been happening.
  
  I can visit C9 with IE6,7 and 8 up to B2 without being told that the site needs the FlashPlayer to be enabled.
  
  Suddenly, with RC1, the site needs FlashPlayer ??
  
  Pretty hard to believe that it's the site that is causing the problem, when it's only just started with RC1.
  
  I have the C9 front page on 2 screens in front of me right now... one running IE8-B2 and the other running IE8-RC1... RC1 wants me to enable FlashPlayer, B2 doesn't care about it... and yet the status is the same for both... installed but disabled for all sites.
  
  Personally, I think the issue is something to do with AJAX and something that is being incorrectly interpretted by IE8-RC1, as all the sites I've hit the "false flashplayer request" problem with so far (including my own) are AJAX sites.

• 

  Dean and Jason should not downplay performance. It is important. I like my page to render fast. More and more website are using pretty neat javascript. Live Maps and Google Maps for instance. Zooming and panning loads multiple images, and I want it to work in a very fast manner. I suspect Google did street view in Flash because javascript in browser just simply isn't fast enough. Chrome really makes live maps fly for me, as does Firefox 3.1 Beta. IE8 is not horrible, but it isn't up to par with the competition. This is "real life" perception differences. I do not care about numbers, I want my browser to load quickly and scroll smoothly and quickly.
  
  Another example is, scrolling on www.anandtech.com in IE8 is kind of slow at the top with the flash banners, then speeds up as they dissappear off screen. This does not happen in Firefox. I will have to post video of this, as no one will probably believe me. It is not my computer either, 4GB ram and a QX9650 and all...
  
  I would love to use IE8 as my main browser again, but if sites are going to be noticably choppy to me when scrolling, I simply cannot deal with it when the alternatives provide better performance.
  
  
  Can any IE7 user print http://en.wikipedia.org/wiki/Proton_NMR with a normal second and first page?
  
  
  
  

• 

  are you running the rc? i had real pref trouble with b2 but the rc is much faster.. if you havent, try the rc and see if it works better

• 

  Nope, the second page doesn't print here too. It works in the IE version that comes with Win7 so they actually broke the printing for RC1

• 

  That is what I thought, my 7 machine works fine as well. I finally found a regression!
  
  @ al_, I am using the RC.

• 

  @RoyalSchrubber: i wasn't able to reproduce the bug you mentioned. i'm running IE8 RC1 on WinXP SP3.
  
  one can see a lot of work going into IE8. 
  btw, how would you guys respond to crashie.com?

• 

  then i got no clue, sorry :/
  
  @Murtaza
  crachie doesnt crach my ie at least

• 

  really!!?? were you running IE in No Add-ons mode?

• 

  it seems ie8 (like its predecessors) has no support for svg, mathml or the canvas tag.
  this is really unfortunate and simply holds innovation back.
  
  firefox has made such major advances - witness the plethora of add-ins, themes, debuggers.
  
  i happen to be a ten-year veteran of microsoft, so this is not by any means a rant. but i stopped using ie for browsing or for development years ago, and see no reason to switch back.
  
  
  

• 

  So IE8 now officially in released crap phase, I have to wonder what on earth I was thinking installing this and going through a system restart for it (When will IE become less entrenched within the Windows OS !!!, I should have known nothing would have really improved since its last beta ...I mean royalscrubbers first comment made laugh just that it was him actually complaining about one of many retarded things to plague the IE browser for years & years.. but your response "We're looking into the issue now. Thanks for pointing it out!" ..what only.. now?

   .. I don't think C9 has/DISPLAYS emoticons that could have shown my expression on reading that response.

  What about downloading files with IE and on the miss hap that the file isn't fully downloaded and is left corrupted in the IE temp cache.. even as quick work around of restarting the download and renaming the saved targets filename, IE will still get the cached version and copy that to the new location/file-rename.. lol so stupid, having to manually clear browser cache just to get something like that to work. It would make more sense that if the user is wanting to save the file out to a different name it shouldn't use the existing cached file regardless of its filename being the same! Its just stupid by design, but then a lot about Internet Explorer seems to have been left that way for so long or developed at the bare minimum, many other areas of IE downloading that should be improved like file resume support if a cached version wasn't finished, or actually just saving the file directly to its targeted location and not to the cache +then waiting for copying process, a mini download manager isn't required(it would be good) but at least do something to improve things.

  And the InPrivate browsing is a bit of a primitive feature its major weakness being a user will always have to go through the same login steps, maybe someday when IE doesn't suck, you might also provide the option of allowing encrypted user profiles (inc temp /cookie data etc) with the option of providing online roaming profile support(perhaps using Live) to keep a users browser settings/favs/rss feeds/addons/skins(perhaps when IE GUI sucks less)  as done in Maxthon2 (really feel sorry for that browser . its biggest downfall was being built to use the IE engine they should have dumped it years ago and moved on during M2's development tbh.

  I guess the only reason I am even bothering to write here is that my default browser is Maxthon and it does unfortunately use the IE engine, despite having Firefox +bucket load of extensions to achieve the same+ more power features I'm accustomed to, it doesn't provide the same experience (better Maxthon addons/ skins) or overall speeds in general workflow usage. I guess I was hoping IE8 engine might have improved general page rendering speeds yet you still lag way behind. You make a point about lab testing performance speeds aren't the deciding factor in a users browser choice (and you're right they aren't, afteral I would dumped Maxthon and this IE engine years ago as I'm sure most web developers would like too as well ) yet after saying that the experience of using IE8/engine isn't better or faster at all, I don't need to see script/css etc benchmarks to feel that this browser still isn't performing as good as other browser engines. Also while I won't attest to using the IE8 browser all that much(its rubbish in so many other ways that even the small improvements like the floating web accelerator and improved address bar features don't at all make a difference) And the delay on opening even a blank tab is silly (even when set to about:blank).. notice the delay when opening a blank tab and it says "Connecting.." lol, does it not know its opening empty tab! (oh it’s an extra useless iexploder.exe being created, great! IE8 GUI and tab support is just terrible, its slower and worse than FF +the many tab addons, even Maxthon classic/2 implements a better tab support in combined features/options and speed using the ie engine, not to mention the IE8 tab options are just plain weak! where are the options to remove the useless [x] close button, I'd imagine a lot of users have mice with middle mouse buttons why not provide a tip, its generally used by all other tabbed document GUIs for closing tabs!... and while i'm on this point as IE is related to how Win7 taskbar usage of the [MMB] on its new crap default iconized noob bar as a way of creating an empty app document/instance. Which I think is just STUPID! its pretty much a standardized thing that in any MDI application with doc/ tab support, has implemented the [MMB] for closing, it’s perfectly suited as a universal option for closing running tasks on the taskbar(especially now that the [RMB] Context menu on the taskbar no longer appears directly next to the mouse position, but instead positioned somewhere above the taskbar.. not only looks crap, workflow wise it’s also lame(like a lot Vista/7 changes !not improvements) ..And how many programs/apps that display on the taskbar would make use of a [MMB] for creating a new document/instance, SOME... Vs. how many would all benefit from a quicker closing down method using the [MMB], the answer is ALL!. Heck there was even demand for an addon for the implementing said feature on the taskbar(taskix), yet you Micronubs go and implement the exact opposite functionality while still as usual providing some piss poor taskbar customizability, let alone anything else that could have been developed like mousewheel support on the taskbar as quicker alternative to use the keyboard methods of switching applications or quicker toggling for that matter!.. but instead we get noob crap!, thanks i'm sure Win7 will be raging success NOT.

  Anyway back to IE8,  newly created tabs appear at the end of the tab bar?-that is stupid, especially when there is no option to have newly created tabs opened adjacent/right side of the currently opened tab... Session restore (this features implementation just makes me laugh compared to Maxthon's out of all browsers.) And Automatic Crash Recovery... this feature alone says it all about IE, the very fact it is mentioned as feature is shameful, you should be designing browser engine to avoid crashing in the first place (crashie.com just makes me laugh), and if a tab did crash the user should be notified about whether they want the same site url to be opened again!! pff afteral it did just cause a crash! Also multiple iexplorer.exe tab processes in the taskmanager is just annoying, need I just point at the chrome browser for how it should have been done tbh, apart from the main reason being its impossible to see visually (without testing and checking cpu usage,) in Windows still barely improving crap taskmanager which iexplorer.exe process belongs to which tab document.

  And why has the Autoscroll feature suddenly becoming an autoscroll_JUMP feature, where it jumps you to the bottom or top of page, you barely have to move the mouse position away ~170px away(that’s NOTHING!)....  its about the exact distance I would naturally move my mouse position away from the initial invoked location in an instant of using the [MMB] only now it skips everything on the page, unless you keep within those bounds and put up with a terribly slow scrolling speed(might aswel just forgot autoscrolling and use the mousewheel), you could say desktop resolution(large) and mouse speed effect this and its a per user preference, which means if there was any sense in the implementation of this change, you would have provided the ability to adjust its settings? a reg tweak even? I just hope it isn't a hard coded setting, doing things like that really piss me off. 

  I'll let you know that out of all other browsers that have implemented autoscroll, IE has surprisingly been the one I have always preferred(perhaps because its always been the same for years & years, and as user of the IE engine have become accustomed to its scrolling speed/mouse distance, beats using the mousewheel when you're just wanting to skim through,)  its also the most directly input responsive autoscroll feature be it in sticky mode or held down,  so why after years & years & years & years of leaving this feature alone do you go and make a change like this? Its just one more thing to put me off the engine now. I mean you slack on not improving the poor autoscroll gfx
  
   
  
  (a minor detail, but coupled with everything else also says alot about the LACK of polish IE really has)  but to change what might appear to be such a small detail on how it functions. And I know IE has the crappest page rendering performance out of all browsers you only have to use the autoscroll feature to notice the jitteriness(specially scrolling past animated objects(that also stop animating when scrolling, just compare that to other engines), large resized images! transparency effects, or just going through heavily designed/content loaded pages etc),sloww... but skipping the entire page contents for a feature that’s meant to scroll past! only a noob would think of that as useful, for anyone else who knew they wanted to get to the bottom of the page contents, would use mouse gesturing / shift clicking the scrollbar / using the [home]/[end] keys all of which are  better suited or just simply scrolling faster!. Please tell me you didn't implement this autoscroll page jumping, as way of saving face on what is still embarrassing page rendering speeds, coupled with page/css compatibility issues, scripting performance, and a reluctance to further implement up and coming specs. I guess the question is when will IE ever catch up, IE9?

  
  Well thats my harsh words over there are good signs of improvement in IE8 -features & dev tools, just overal not enough at all.

• 

  That autoscroll icon does bother me as well. In addition there are some nice advanced features in Firefox that I use on a daily or weekly basis that prove to be useful.
  
  Take for instance this page. http://science.howstuffworks.com/oil-refining4.htm
  
  See that beautiful flash animation, I want to see it bigger. I use Firefox's page info to find the direct link put it in the address bar and it is now blown up.
  
  
  
  
  
  In IE, the only way I can do that is view the source, and assuming I know what the extension is, search for that.
  
  
  
  The biggest advantage to Firefox is I can explain to someone how to do that easier in Firefox than in IE. I always get asked "how did you ______".

• 

  I have been giving feedback to the IE team on the autoscroll graphic and advanced features for a long time, but obviously nothing has changed...
  
  Well I think it did, if you look at my pervious post on this, the icon had no black border.
  
  https://channel9.msdn.com/forums/Coffeehouse/92220-Someone-who-beta-tests-IE7-I-have-a-questoin-for-you/https://channel9.msdn.com/forums/Coffeehouse/92220-Someone-who-beta-tests-IE7-I-have-a-questoin-for-you/
  
  
  
  I love IE's process separation, I love IE's favorite system, I love IE's RSS feed handling and I love IE's zoom. There are good things the IE team is doing, but there is so much left to be desired.
  
  Edit: I could not edit this post in IE 8 RC1, it just failed to post with an error, this had to be done via Firefox.
  
  I've restarted the browser, cleared data files...etc. I get this permanent "saving".
  
  
  
  Webpage error details
  
  User Agent: Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.0; WOW64; Trident/4.0; SLCC1; .NET CLR 2.0.50727; .NET CLR 3.0.04506; Media Center PC 5.0; InfoPath.2)
  Timestamp: Fri, 30 Jan 2009 04:20:58 UTC
  
  
  Message: Object doesn't support this property or method
  Line: 68
  Char: 12094
  Code: 0
  URI: http://channel9.msdn.com/Telerik.Web.UI.WebResource.axd?_TSM_HiddenField_=ctl00_RadScriptManager1_HiddenField&compress=1&_TSM_CombinedScripts_=%3b%3bChannel9.Web.UI%3aen-US%3acafd9df3-245b-4200-b60a-9e6b4426d322%3a33b5aa35%3ab1f64471%3ab6ad9aee%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3ad0cc46e9%3bChannel9.Web.UI%3aen-US%3acafd9df3-245b-4200-b60a-9e6b4426d322%3a1b1af369%3bChannel9.Web.UI%3aen-US%3acafd9df3-245b-4200-b60a-9e6b4426d322%3abfc6b98d%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3ac32548cc%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3ac0777a51%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3a78fefdd1%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3a3756f318%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3ab7126276%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3add268e21%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3a3ca51081%3bChannel9.Web.UI%3aen-US%3acafd9df3-245b-4200-b60a-9e6b4426d322%3a46ab4dab%3bChannel9.Web.UI%3aen-US%3acafd9df3-245b-4200-b60a-9e6b4426d322%3a8baa875a%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3af1ce222a%3bChannel9.Web.UI%3aen-US%3acafd9df3-245b-4200-b60a-9e6b4426d322%3a7c64497d%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3a974f2aad%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3a52d56b39%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3a5db195a6%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3a1bfc391c%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3ae5cfa4cd%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3abc8b7265%3bEvNet%3aen-US%3a09cd26da-04ed-4c36-b7e2-46986a4a3dff%3a6848da3d%3bTelerik.Web.UI%2c+Version%3d2008.2.723.35%2c+Culture%3dneutral%2c+PublicKeyToken%3d121fae78165ba3d4%3aen-US%3afaa3c25c-0b63-45d2-9c6e-561f75337212%3afe8d4455%3bTelerik.Web.UI%2c+Version%3d2008.2.723.35%2c+Culture%3dneutral%2c+PublicKeyToken%3d121fae78165ba3d4%3aen-US%3afaa3c25c-0b63-45d2-9c6e-561f75337212%3ac7991a52
  
  
  
  
  

• 

  I have noticed one more major bug. I'm doing lot of Flex and Silverlight development. The problem exists when running IE8RC1 as default browser for Flex Builder debugger. When closing IE8 the process stays running almost all the time. Processor usage is 0% but message usage goes up to 155MB/process. Sometimes I notice I have like 8 of those hanging in memory.
  This happens on Vista Ultimate 32-bit.
  

• 

  This is the last time I'm going to try to explain this to you.

  This page tries to instantiate Adobe Flash.  IE 8 RC1 detects that and assumes that this page needs Flash.  It looks in its set of sites that are allowed to instantiate the Adobe Flash ActiveX control and notices that C9 is not one of those.  It tells you so and asks if you would like to add C9 to the list of sites that can instantiate the Adobe Flash ActiveX control.

  The default policy for which sites are allowed to instantiate which ActiveX controls may have become stricter between IE 8 Beta 2 and IE 8 RC1. Or perhaps IE 8 RC1 just forgot all those permissions you set in IE 8 Beta 2, because the internal data structures used to track such information changed, and they didn't bother to write code to migrate those settings from the Beta 2 format to the RC 1 format.

  Click Tools | Manage Add-ons, select the Shockwave Flash ActiveX object, click the More information link, and see which sites have been approved to run the ActiveX object.

  That's how it works. There's no mystical "Ajax sites cause Flash not work in IE8" BS going on, ok?  Many sites use some Web analytics software (like Microsoft uses WebTrends) that, as part of their analysis, has script to detect which version of Flash your browser has installed. That's what's causing this problem, not a site being an "Ajax site."