Does MS internal email confirm Explorer can be removed?
It certainly seems to contain a useful list of which files do what - can we all have it now please?
The US Department of Justice yesterday filed an emergency motion demanding that Microsoft turn over an internal email which it says confirms that Windows 98's Internet and shell functions can be split. According to the DoJ, the email agrees with claims by prosecution witness Edward Felten that it would be possible to pull apart shared DLLs, and thus remove Explorer. This is not exactly big news for the programming world, but a Microsoft admission that at least some of Felten's claims are true would undermine counter-claims that have been made by MS execs Paul Maritz and Jim Allchin. The latter's deposition, which was made public yesterday (Browser integration predates Gates' birth), contains great swathes of data/invective on the inseparability of the two. The Microsoft email was sent during the period when Microsoft was testing Felten's prototype IE uninstall program for Windows 98. There has already been some argument over this period, as the DoJ claimed that Microsoft had completed testing but avoided telling the DoJ about the results. Felten's program seemed to get accidentally broken just before he was due to take the stand. According to the DoJ motion of yesterday, the email says "Arguably, based on Felten's testimony, this list [of shared files] could be used to 'separate' shdocvw into two parts: shared+shell and browser specific." This is not of course rocket science: shdocvw.dll is tagged by Microsoft as a "Shell Doc Object and Control Library," and it's in the nature of DLLs that they include functionality covering more than one program. Or as Microsoft might put it, 'more than one set of technologies.' If it's simple to pull the functionality apart, of course, then that's helpful to the DoJ's case - it might suggest the components were arbitrarily put together in the first place. And if you could do it without breaking anything, well, that helps the DoJ's case some more. Microsoft was last night quoted as saying the email showed that there was an enormous overlap in browser and shell files in Windows 98, but this makes it interesting too. If the email contains a list of overlaps, then it would be generally instructive, and highly useful for people working on pulling Explorer out of Windows 98 (Example well worth visiting). According to Allchin's deposition: "Whether a piece of software or some functionality can be 'removed' from a software product does not mean it was not a useful, integrated part of the product line in the first place. My hand can be surgically removed from my body, but it was certainly a well-integrated part of my body before that surgery." Spookily, we heard somebody else from MS say that very thing some weeks back - boy, are these guys on-message. "The central issue should be whether customers benefit from an integrated design, not whether such a design can be torn apart later." That's certainly one of the issues, but not necessarily the central one. Here's a possible alternative candidate: "They [browser users surveyed by MS] said they would not switch, would not want to download IE 4 to replace their Navigator browser. However, once everything is in the OS and right there, integrated into the OS, 'in their face' so to speak, then they said they would use it because there would be no more need to use something 'separate.' The stunning insight is this: To make them switch away from Netscape, we need to make them to upgrade to Memphis." That email and the associated survey data were sent in February 1997 to Bill Gates, Paul Maritz and Jim Allchin. So what do you think it means? It would appear to be saying that the survey data clearly indicates that by integrating in Memphis/Windows 98 Microsoft can cut off Netscape's air supply. The central issue therefore could be whether this was the primary objective, and the integration was carried out for that reason, or whether the eclipse of Netscape would merely be collateral damage. Allchin also says that the core IE files remain in Windows 98 after Felten's uninstall, and that "The DOJ version of Windows 98 is 99.93 percent as large as the Microsoft version of Windows 98." Again, this is pretty meaningless stuff, as Felten was demonstrating that Windows 98 could work without Explorer - do that successfully, then you can start optimising to reclaim disk real estate. His claims that Felten's uninstall damages Windows 98's performance are however potentially more serious. After you've run it he says: "Various features of the operating system are inoperable," but it's not clear what these are. The 98lite site we referenced earlier in this story doesn't seem to agree. "Any attempted installation of Internet Explorer 4.0 will malfunction" - This might make you happy, rather than sad. "The performance of Windows is reduced" - again, 98lite claims the reverse. " It causes numerous third party software applications to fail" - again unspecified, but here's a suggestion. The ThinkPad 600 which is producing this very story came with both Netscape and Explorer 4.0, and with a CD of bundled software you can use to add and remove applications. But it's hard-wired to use IE 4.0 as the UI for doing so. Open Netscape, try to open the HTML files to run the add/remove routine, and it'll tell you the program needs Internet Explorer. Did IBM think of this all by itself, or is it in an MS OEM contract? ® Complete Register trial coverage
Sponsored: Benefits from the lessons learned in HPC