toastytech.com — An interesting look at what eventually became Windows 95. "Chicago is unique as it was really the only time in Windows history that Microsoft seems to have put significant effort and design research in to the Windows user interface. "
Feb 5, 2006 View in Crawl 4
rosewoodFeb 6, 2006
Well, considering MS has spent quite a bit of money giving grants to human factors programs all over the country, including the one at Wichita State University, for hardware and software design studies I have to say "No" to the author of this so called article.Maybe one day on digg an article will get posted by the first and only guy to write an article like this and do actual research!
mschaefFeb 6, 2006
"Seems that there is a distinct crossover between Win95 and Win3.11 in this build. The last screenshot seems to exhibit elements of both."So does the internal architecture. The guts of Windows 95 are a slightly evolved version of the combination of DOS 6.22, Windows for Workgroups 3.11 and the Win32s extension. There was a lot of publicity about it being a 'complete 32-bit rewrite', but that was really the Windows NT kernel, which had problems with running on small machines and backwards compatibility with all the funky DOS/16-bit windows stuff that folks had been developing since 1985.If you're interested in the details of the timeline, here are when key bits of functionality were released:Windows 1.0/2.0 - 16-bit real mode Windows APIWindoes/386 (~1988) - The Windows API was still 16-bit real mode, but it ran on a 32-bit VxD layer. This is the first version of Windows that could natively run 32-bit code, but you had to write a device driver to do it.Windows 3.0 (1990) - The 32-bit VxD layer is still around, but now the Win16 API can optionallly run in protected mode. This gives Windows apps the ability to access lots more memory. Windows 3.0 has the ability to fall back to entirely real-mode operation on limited hardware.Windows 3.1 (1991-2) - Real mode operation disappears.Windows for Workgroups 3.11 - Since Windows/386, the VxD layer is given more and more responsibility. WfWG 3.11 (not WfWG 3.10) contains a 32-bit implementaiton of hte FAT file system. This code, a precursor to the long file name code in Win95 allows Windows apps to completely bypass DOS when accessing the disk.Win32s - This is a layer atop Windows 3.1 that enables it to run a limited set of Win32 apps. It hooks into the Win16 layer and adds a function, ExecPE, that's a Win32 executtable loader. The Win32 DLL's in this environment are almost entirely composed of 'thunks' that translate Win32 calls into their Win16 equivalents. Windows 95 - This version extends the 32-bit FAT code to support long file names. It also starts implementing some lmited set of the Win32s calls using 32-bit code in the VxD layer, but the fundamental pieces of the architecture are the same as before.There's a lot of mythology around this set of kernels, but even as late as Windows ME, DOS and 16-bit Windows code played an enormous role in the basic operation of the OS.
Closed AccountFeb 6, 2006
I defecated on my windows CDs.
barnstormerFeb 6, 2006
I got to play with beta that summer. Although 9x had a lot of good things going for it, I was amused at how many UI defaults I had to change to get a useable system. Some of them are still defaults today.
barnstormerFeb 6, 2006
That summer? Obviously I meant the next year. I don't remember a lot of the early nineties.
blasterman95Dec 6, 2007
Sintax, Snowball was Windows For Workgroups 3.11.Toastytech is an awesome site.