A peek in to Cairo and how Microsoft plans Windows

It’s funny how history has a way of repeating itself when it comes to overly-ambitious Microsoft operating system projects. A once confidential Microsoft document released as evidence in the Comes V. Microsoft lawsuit gives us a rare insider look at how Microsoft plans for a product as complex as Windows.

Although I have to admit I have not yet built up the courage to read this 76-page product planning document for Windows Cairo from 1993 word for word, but even skimming this document has revealed a number of interesting tidbits that shows just how forward-looking this project must have been for its time.

One example that stands out from the rest, the document describes the concepts of “Smart Folders” which presents folders to users that both store and view information based on metadata as well as “finding information through queries” – ideas later reanimated as part of the Longhorn WinFS vision but still not fully realized 17 years later even though the problem has been and is still relevant today.

The original case exhibit PDF is downloadable for bedside reading here.

11 insightful thoughts

  1. Interesting to see the current president of the Server and Tools business, Bob Muglia in there as a feature owner 🙂

  2. I’ve just started a long night of reading-I’m on the 8th page-and a lot of it seems to be poorly implemented in Windows Vista.

    1. On page 11, there’s mention that it won’t support OMG COBRA. I found this both entertaining and puzzling. What is OMG COBRA?

  3. reading this is a bit upsetting and makes me think of all those “what could have been” scenarios, forward thinking? YES!, ahead of its time?…DEFINITELY!

    but what has all the money and time amounted to..there is no Cairo and this goes all the way back to 1991 or so, they couldnt even give it a defining name of being an OS or a “collection of technologies”, its funny because even i get excited when i read things like this, but then it makes you think about how “forward thinking” they were back then and compare that to what we have on our desktops today…Windows 7 seems barely ahead of its time, although i have no complaints about the OS as a whole..just seems like we got the shaft thats all…

  4. I’d love someone to transcribe that…I can barely read the poor quality scan, which is a shame.

  5. I glanced at the plans. What a mess!!! It’s a bunch of corporate-speak. It keeps referring to other products and how it will extend the features or UI or anything else. It’s like reading government legislation.

    Windows products do work, but that’s the problem. Its hard to make things work and how to access the features and benefits. They are so far behind in terms of consumer awareness. People use it, but don’t know how to extend its features. They will always be behind since they rely on their OEMs to exploit the features.

  6. hah these documents haven’t upgraded much over the years. Still at least there is a document outlining the intent you’d be surprised how many products have launched still haven’t gotten this far 🙂

    Speaking from experience


    Scott Barnes.
    Former Product Manager Microsoft.

  7. “CAIRO” was released to RTM labeled as ” Windows NT 4.0 workstation” released in 1996.

    NT 3.5.1 workstation was released only one (!) year before in 95 and was a rock solid OS.

    It had Windows 95 application support but featured the Win3.x UI ( progman.exe ) which made it look “old fashioned”. NT 3.5.1 introduced OpenGL to Microsofts “Business/Professional” Systems ( its 3d screensavers are shipped in every windows, including Vista since then ) and was even supporting a wide variaty of those “designed for windows 95” apps, hell even Office 97, winamp and Mozilla Firefox 2.x run on it.. Outdated UI and missing taskbar was the only drawback for its usability.

    The only visible ( on a average user level ) change of NT 4.0 was the Win95 UI ( explorer.exe ) which brought the 1995 invention of the year – the taskbar and startmenu of windows 95 – to NT users. But besides that very few has changed on the first look. It had the same limitations as 3.51 : no plug and play, needed excessive RAM compared to the “consumer” OS Win95. ( Win95 needs 8MB and NT needed at least double to triple that, for optimal performance around 32 to 64 ), needless to say that NT was only found on expensive professional machines and hardware support / drivers were rare.

    For most sysadmins back then there was no compelling reason in the first place to upgrade from 3.5.1 to 4.0. Compare it to Vista (NT6.0) vs “7” ( NT 6.1) : only a minor upgrade,tweaked for performance and a changed shell UI, but more or less the underpinnings are the same. Technically it does not matter if you run vista or 7. And that was the same for NT 3.5.1 versus NT 4.0. Microsoft even made the same mistake back in 1996 as with now : label NTnext as “4.0” instead “3.5.2” only to make it sound as if it was a major realese which it was not. Win7 is not NT7 but 6.1, i am sure you all know this by now and also that “7” is only a very small upgrade to “Vista”

    So just compare the final product NT4 with the outlining ideas in that “CAIRO” document to see what has been dropped/changed/added.. Without reading it in full I am sure Microsoft made the same mistakes as with its Longhorn “visions” : promise and never deliver.

    Vista is a joke compared to what I have seen and used on my MSDN testmachines back in 2004 when I played with “Longhorn” and explored the wonders of “WinFS” “Avalon” and “indigo”. Who could know that this was only a techdemo that would never come out in the final product. Who would know that applications I had developed on Longhorn would never run on the final NETFX 3 once it was baked into Vista ( they made the final product incompatible to its own alphas / betas before ) .

    Making promises is another of Microsofts all time mistakes. They simply should talk only then when it is ready to showcase and a production ( rtm ) test has been succesfully evaluted internally.

    BTW : I am running Vista ( wont upgrade, read above posting to understand why.. ) and If I just for a minute think away the UI of it, think of the resource usage and all the “negative” sides of it I have to say : It is still NT. Microsoft is good in adding polish to it, extending its kernel to support usb,ide,s-ata,blueray etc etc. But really. Its the same workhorse from back then, no matter if they brand it “2000”,”XP”,”Vista” or “7”. even “8” will be the same.

Comments are closed.