Links to exhibits from the Microsoft anti-trust case, with a few notes on some of the links. There are 945 public exhibits (the highest numbered exhibit is 2519, but many exhibits were only admitted under seal and many were also not admitted), so this is a work in progress as I read through exhibits.
I think it's striking to compare pop conceptions of what Microsoft execs were thinking vs. actual thoughts from execs, e.g., Paul Graham, writing almost a decade after these documents went into the public record, in typical comment about Microsoft cluelessness, said
Gmail also showed how much you could do with web-based software, if you took advantage of what later came to be called "Ajax." And that was the second cause of Microsoft's death: everyone can see the desktop is over. It now seems inevitable that applications will live on the web—not just email, but everything, right up to Photoshop. Even Microsoft sees that now.
Ironically, Microsoft unintentionally helped create Ajax. The x in Ajax is from the XMLHttpRequest object, which lets the browser communicate with the server in the background while displaying a page. (Originally the only way to communicate with the server was to ask for a new page.) XMLHttpRequest was created by Microsoft in the late 90s because they needed it for Outlook. What they didn't realize was that it would be useful to a lot of other people too—in fact, to anyone who wanted to make web apps work like desktop ones.
But if we look at what execs said in the mid 90s, which encompasses a lot of thoughts running into the early 90s, Microsoft execs knew exactly what they were doing they extended web standards and had fairly good (as in business savvy, not moral) reasons for doing so. Their game plan didn't work out for a variety of reasons, but the fact that they managed to half execute their strategy is the reason a $2T company today in 2022 (the 2nd most valuable tech company on the planet), while the traditional competitors they were worried about and discussed at length (Novell, IBM, AT&T, Yahoo, Sun, etc.) are now dead or irrelevant and Microsoft is the company that created TypeScript, LINQ, VSCode, WSL, etc.
There are real reasons that Microsoft's strategy didn't work and lessons to be learned from those reasons, but learning those lessons requires actually understanding what happened and not just trashing Microsoft as a company full of bad leadership that doesn't have any good programmers.
- Projection of PC marketshare
- Share of new browser users
- Share of the browser market, grouped by major ISP, 3 month moving average
- Share of the browser market, grouped by major ISP, monthly average
- IE vs. Netscape marketshare
- IE vs. Netscape marketshare, table with incremental numbers
- IE vs. Netscape marketshare, table with incremental numbers
- IE vs. Netscape marketshare, table with incremental numbers
- Netscape quarterly revenue from browser licenses
- Netscape quarterly revenue from browser licenses without subscriptions for upgrades
- Comparison of MS share of browser shipments depending on whether or not IE was the default browser for the ISP
- Contractural restrictions on preferred browsers at the top 80 ISPs, December 1997
- Examples of OSes that Netscape runs on
- Projected IE marketshare (with low, medium, and high outcomes)
- Projected IE vs. Netscape marketshare
- Multiple emails about Netscape and the internet, emails are in reverse order
- RE: Internet as a business tool email from Pat Ferrel
- Two major concerns about the internet ("the internet is about as open as it gets" and "the internet defines formats and architectures that MS has no control over and very little say in")
- Potential proposals to counter above concerns
- Can create "open" standards that MS controls
- Can own the browser space, as evidenced by "how quickly Netscape took over from Mosaic", giving MS control over formats and architecture
- Can build MS ecosystem to "entice people from supporting non-MS architectures like URLs, WWW, Acrobat ..."
- Solutions above not ideal, but must do something before Netscape owns the space and defines how the web works
- RE: Internet as a business tool email from Russell Siegelman
- Agree with [some previous email; unclear if it's the one above in the exhibit or some other email that wasn't included], but concerned about underestimating "publisher/ISV threat"
- Concerned that API hooks into Netscape are effectively a new client-server platform
- MS has more resources, may be able to "head them off" if action is taken immediately
- Would be mistake to fight Netscape with just a better browser; should use integration with [various products]
- Some kind of memo that describes Netscape's offerings, an attachment to email below
- RE: Internet as a business tool email from Paul Maritz
- Need to meet to come "to get consensus on our basic approach to the Internet"
- Internet as a business tool email from Bill Gates
- "I know I am a broken record on this but I think our plans continue to underestimate the importance of an OPEN unified tools approach for the internet. The demo I saw today when Windows 95 was showing its Internet capability was someone calling up the Fedex page on the Internet and typing in a package number and getting the status. Imagine how much work it would have been for fedex to call us up and get that running on MSN and negotiate with us. Instead they just set it up. A very simple way to reach out to their customers ..."
- RE: Internet as a business tool email from Pat Ferrel
- Internet email thread; emails are in reverse order
- Re: Internet (aka Web Windows) email from johnlu@
- Remoting to another machine and conferencing; web not 1:1, 1:1 conferences "aren't going to turn the internet on its head", should focus on things that have mass appeal and think about 1:many scenarios
- FW: Internet (aka Web Windows) from Paul Maritz, forwarding email by Rick Rashid that forwards another email from Rick Rashid
- Additionally, web windows could provide shared app spaces
- "As an example of Bill's point, Dan Ling showed me this morning that Satan ... actually uses HTML as a user interface. Other software may begin to do this. Its easy. It produces a reasonable looking machine independent UI rather quickly. Its dangerous from our perspective of wanting to make and preserve valuable standards..."
- HTML currently very limiting compared to Citrix, "moral equivalent of an X-Terminal". Citrix can work even over 14.4
- "idea: make Windows the standard interactive application interface for the internet"
- URL could point to spawned application
- Can play back with a browser on Mac or Unix or directly with GDI on Windows
- Would encourage developers to make Windows apps
- Internet email by Bill Gates
- "Given that we are looking at the Internet destroying our position as the setter of standards and APIs do you see things we should be doing to use ACT assets to avoid this? I admit I find it hard to focus lots of resources on trials and things when the Internet is taking away our power every day and will have eroded it irretrievably by the time broadband is pervasive on the course we are on right now"
- Re: Internet (aka Web Windows) email from johnlu@
- Netscape as a merchant channel email thread
- RE: Netscape as a merchant channel email from Peter Neupert
- Need a strategy instead of playing catch up; can deal with Netscape after we have a strategy
- At a minimum, want STT to be payment industry standard over SSL or SHTTP or something proposed by ATT or AOL
- Getting Netscape to put STT into their merchant servers is good; can differentiate elsewhere
- RE: Netscape as a merchant channel email from Dan Rosen
- Agree that lots of choices soon. Acceleration of Internet related activities makes it critical to make choices ASAP
- Should strive for e-commerce market leadership
- Can do it by ourselves, but risky and currently underresourced
- If we partner, should get something back. For example, with UUNET, we maintained account control. Problem is we don't know what we want
- Netscape is a small FOCUSED competitor or ally
- Need to know if, as ally, they'll cede client and its standards to us, etc.
- If competitor, we'll fragment the market and force Netscape into an alliance with someone we don't like, maybe Lotus or AT&T. "I fear the terms of engagement will be set by someone other than us. This is not a path to victory"
- Need to chart a course. Would be easier if we could organizationally align the Internet assets within Microsoft. I not, need strategic alignment on direction, which doesn't exist
- RE: Netscape as a merchant channel email from Russel Siegelman
- Need to think carefully about what to do with Netscape
- There's a feeling we need to compete. If we compete, may not want to license Netscape some technologies or otherwise support their merchant strategy
- OTOH, it may help us if STT is supported everywhere
- Should think carefully before deciding on this
- Netscape as a merchant channel email from Dan Rosen
- Talked about potential structure of relationship with Netscape
- MS will include STT into O'Hare client along with SSL for e-commerce transactions
- MS should try to get Netscape to put SST into their server products and have them adopt our formats
- Time to market is very important and better technology might lose if a competitor (IBM, AT&T, MCI, AOL, and others) gets to market earlier
- RE: Netscape as a merchant channel email from Peter Neupert
- Exhibit 19 not admitted
- The Internet Tidal Wave memo from Bill Gates (May 26, 1995)
- Vision for last 20 years: exponential improvements in computer capabilities would make software valuable. In next 20, communication network improvements will dominate computer power improvements. This will fundamentally change work, learning, and play. Variety and volume of software will increase
- Most users haven't seen comms cost decline significantly due to cable and phone networks depreciating old tech, service monopolies and other govt involvement keeping prices high. Private networks and internet will change this. PC is creating demand which will drive investment. Broadband networks justified by video, etc.; will see low cost communication into most business and homes within next decade
- Internet developments over next few years will set the course of the industry. The internet is very important. Crucial to every part of our business
- Single most important development since development of IBM PC in 1981. More important than GUI
- PC analogy is apt. PC was initially poor in many ways but created phenomenon around it that made it key for next 15 years.
- Companies that tried to fight PC standard often had good reasons but nevertheless failed because phenomenon overcame any weaknesses that resisters identified
- HTTP over TCP is the future. Other protocols will be used, but HTML with extensions is how information will be presented; various extensions (tables, secure transactions, etc. will be widely adopted soon and will have 3D later
- Internet buys communication lines on a commodity bid basis; only "public" network whose economics reflects latest advancements in communication technology
- Price paid is by size of pipe not usage, making marginal cost of extra use "essentially zero", encouraging heavy use
- Bootstrapped into positive feedback loop: more users -> more content -> more users
- Everyone on exec staff and their directs should use the internet
- Try out the following sites, attached via HTM or use Word
- YAHOO is of particular interest
- SUN, Netscape, Lotus present their products well on their websites
- Easier to find information on web than on MS corp network
- Inversion where public network solves a problem better than private network
- Goal for Office and Systems is to allow publishing on LAN. Can be leveraged into HTTP/Web
- Critical issue of runtime/browser size
- Can superset web with Office, but need comparable performance to web with our extensions
- Most important element of Office 96 and next major release of Windows
- real time audio and video content a technical challenge
- packet network doesn't guarantee rates; congestion determines throughput
- [summary of current state-of-the-art approaches, academic as well as commercial]
- All of these will improve as Internet gets faster and software improves
- ATM backbone will have QoS in next few years
- Need improvements to deliver full promise of Internet, but improvements will definitely happen
- "virtually every PC will be used to connect to the Internet" and Internet will drive PC sales
- Expect modem use and eventually ISDN for internet as well as point-to-point
- Using voice data while using Internet useful for trip planning, contract discussion, getting tech support, etc.
- Will eventually be able to, while growing a website, press a button and talk to an agent with relevant expertise... via ISDN
- Cable modems will be another way to connect to internet
- Early systems will essentially turn coax into Ethernet so everyone in the neighborhood is on a LAN
- Upstream from PC to cable system will be the most difficult problem
- Downstream data rates better than ISDN
- Cable providers will eventually upgrade to ATM-based system using all fiber or combination of fiber and coax
- Unclear when this will happen. If soon, relationship between broadband and Internet will be loose; if later, these systems will be an extension of the Internet; later seems very likely
- Big 3 developments of last 5 years are CDs, growth of online usage, growth of Internet
- These are strongly related and growth will accelerate as these come together
- Example: "every online service has to simple be a place on the Internet with extra value added"
- Don't have a clear answer to why publishers or users should use MSN instead of the Internet
- Will need to make MSN very cheap or maybe free for users who don't connect to the Internet through us
- A lot of free information on the Internet
- Although there's room to differentiate from free content, difficulty will put pressure to figure out how to get advertiser funding
- CD-ROM businesses drastically impacted by Internet. Example: encyclopedias
- Traditional competitors just getting started with the Internet
- Novell surprisingly absent given importance of networking to Novell
- Good directory service important; Novell working with AT&T to use Network Discovery Service for this; this is a major threat to us
- Lotus already shipping Internotes Web Publisher, which replicates Notes DBs into HTML
- IBM has Internet through OS/2
- UNIX vendors benefitting from the Internet since default server is a UNIX box and not Windows NT
- SUN exploiting this. Many sites, including Paul Allen's ESPNET, put a SUN logo on site in exchange for cheap hardware
- [other SUN initiatives]
- SGI backing VRML, which will allow VR shopping, gaming, and socializing
- Growing web, find almost no MS file formats; in 10 hours, didn't see a single Word .DOC, .AVI, or Windows .EXE other than content viewers or other MS file format
- Lots of Quicktime; all movie trailers in Quicktime
- Apple benefited from having early TCP support and its building a browser from OpenDoc
- Apple will push for OpenDoc protocols on the Internet
- PDF also popular; HTML limitations make it impossible to provide rich layout; PDF now alternative standard
- Quicktime and PDF popular because cross platform and free
- New competitor is Netscape. Dominant browser. 70% share; gives them control over what network extensions will catch on
- Possibility of a machine cheaper than a PC but powerful enough for web browsing is scary
- Many groups already working on this, but need to have unified strategy
- Need to improve Microsoft homepage
- Internet will become our most important promotional vehicle; paying people to link to our website a good way to spend ad dollars
- Any info we create should be done on our internet server
- Need to make NT highest performance HTTP server [various technical details on parts of this]
- Need a client with integration and need to work with OEMs to ship our browser preinstalled to get people to switch from Netscape
- Shell and browser will converge; need to establish OLE protocols as way to share rich docs on web
- Need to give away client code that encourages used of Windows-specific file formats on web
- Need to make it easy to design a form that can be HTML. CGI is used for "forms" behaviour but is hard
- Need an integrated search engine strategy
- Everything should be output as HTML as well as our extended formats
- Where's our competitor to PDF? It's inadequate
- Critical actions
- Office needs compatibility for Mac and Windows
- Merger of online business and Internet business presents a challenge for MSN. MSN can't just be the place to go to find MS info on the Internet because much of the content we're getting will be surpassed by what's on the Internet * Consumer org. Online is great for annuity revenue and limited shelf space for titles, but there will be a large number of free titles. Competition brutal compared to today's CD market. Need to decide if we'll be #1 or #2 for each category or shut down business
- Need a strategy for broadband media applications; need strategies for directory, news, shopping, local info
- E-commerce: online gives us chance to take a new approach to challenge Intuit and others. Need to figure out how to enhance Money. Maybe a tax business that only operates online or lowest cost bill pay or team up with Quickbook competitors. Financial institutions will be able to buy technology and tools to compete here without much technical expertise
- The Future
- Strengths for us are Windows and Office
- Things moving quickly; will have to revise strategies
- The Web is the Next Platform (version 5) memo by Ben Slivka (May 27, 1995)
- Summary
- The Web (aka, the Internet and its data formats and protocols) will grow rapidly and become a platform rival that surpasses Windows
- Main case memo makes
- Web is an app platform that's a threat to Windows. Corp devs and ISVs could developer more quickly and deliver to a wider audience with The Web than with Windows or MSN as they exist today
- To influence the Web, MS "must have broad, standards-based Web support" in products. Need to be "the product supplier of choice for all key existing Web technologies — clients, servers, and publishing tools, at a minimum"
- Once we have market share, can take leadership role in expanding/shaping Web
- [other notable things from this doc: what's wrong with OLE and why MS started taking backwards compatbility seriously]
- Why is Web a threat to Windows?
- Rapidly maturing app delivery platform. Can shop, play games, have discussions, read news, check weather forecasts, stock prices, get dealer costs for cars, download movie trailers, [and many other examples]
- Can call anywhere in the world with an "internet phone" (CB radio quality, but will improve), read USENET (NNTP), chat (IRC), 3d chat (Worlds Away), stream audio (RealAudio, from ex-MS folks)
- Nightmare scenario is Web grows into a rich application platform in OS-neutral way, then Siemens or Matsuhita makes $500 "WebMachine" that attaches to TVs
- WebMachine could let customer do all the Internet stuff with $500 box with RISC CPU, 8Mb RAM, no HD; $2K P6 Windows machine less attractive to 2/3 of homes that don't have PCs
- Example cool stuff on web today
- Virtual Vineyards. See wine list, charts of wine flavors, add wines to basket, order wine. Have session id in redirect URL when you first visit and use this to track state
- Play rubik's cube. Uses a sensitive map. Click on arrows to rotate
- Send a web postcard
- Movie database and rating system. Internet user-maintained movie DB. More movies than Cinemania. Reviews from any random web surfer. 45926 titles, 4569 plot summaries; 79619 actors, 43942 actresses, 10289 directors. Shows use of HTML as UI to DBs
- Why do we need to start from the Web today?
- If we don't act quickly, will lose standard setting role we have with DOS and Windows and Office and the associated profit margins
- Used to think "we could just 'build it and they will come' with MSN"s; protocols like Marvel RPC incompatible with mail and news protocols, etc.
- MSN would be further along if we started with existing Web and enhanced it
- Battle for online world largest challenge MS has ever face and we're late
- With BASIC, almost no competitors. DOS, few small competitors. Windows, maybe half-dozen competitors. For Web, "we have all of our favorite" software cos (Novell, Lotus), hardware (Sun, IBM), new (Netscape), phone (MCI, AT&T), ISPs (PSI, Netcomm), banks, etc.
- UNIX model one possible evolution. Serval companies start with an ideas but go in proprietary directions
- Possible if we pay lip service to the Internet, "we may 'pull a Windows'" and dominate Web; other players fight over standards and incompatible extensions
- But someone else may "pull a Windows" by taking a leadership position and enhancing the Web. Netscape is pursuing this strategy
- Key mistake from past large efforts, OS/2 and LanMan, was not enough focus on compatibility; OS/2 had no DOS compatibility and LanMan no Novell compatibility
- How should we extend the web?
- Support all key standards and extend with MS technologies
- Key Issues
- [lots of technical details on what protocols to support, how those interact with existing MS efforts, risks of "BlackBird" project, etc.]
- Proposal: Microsoft SuperWeb Architecture
- "OLE without the E"
- [quite a few details, including on the advantages of HTML over Word doc files]
- Recommendations
- NT
- Major improvements to scalability and reliability of NT (monitoring, fault tolerance, sharding, etc.). microsoft.com can't use NT replication service "because it doesn't deal with open files"
- VB for servers; need a tool for quick development of webapps. Instead of VB forms, this would create web pages. Integration with publishing environment "critical"; maybe should be BlackBird team instead, using BGI hooks NT has
- Query services. "OFS is its own little world"; can''t scale. Split up storage, indexing, and querying so they can be done on separate machines, like Lycos model today. MSR NLP folks have a way to determine webpage similarity, so can add "find similar pages" button to IE
- etc.
- O'Hare (now known as Internet Explorer)]
- "Break [IE] into: Frame window, HYTML "docObject control", Protocols, Caching"
- Scripting for HTML
- HTML 3.0, esp. tables
- Multimedia embedded
- HTML editing
- SSL, HTTPS
- [many other groups, product lines]
- NT
- Business opportunities on Web (format is Area / Priority / Total Market / Comments)
- Client / 1 / 0 (or small) / client is just to sell Windows; need it to be ubiquitous to control evolution of Web
- Server / 1/ $$ /
- Publishing / 1 / $$$ / "Like a newspaper/magazine publishing system: manage workflow, media, ..."
- eCommerce / 1 / $$$$$... / "Getting STT established ASAP is our only chance"
- Content / 2 / $$$$$ / "have to decide how much publishing vs. creation we do"
- Server Farms / 2 / $$$ / Low margins
- Plumbing [referring to routers, backbone, etc.] / 3 / $$$$$ / Low margins 0
- "Our uncoordinated approach to the internet so far"
- Word views web as info retrieval system, insists HTML be replaced by richer DOC, XLS, etc.
- MOS views web as online service, "insist that the existing 'weak' protocols" and formats be replaced by proprietary protocols and formats (MOSView, MPC, BlackBird; no support for SMTP, NNTP, HTTP, etc.)]
- NT views web as another market to sell NT and BackOffice. Contrast with Netscape's focus (merchant servers, protocol enhancements (SSL), server enhancements (see 5/9/95 press release on Verity search engine and Netscape server))
- O'Hare views Web as new application platform; avoiding OLE, OCX, COM, etc. to get smaller size and more speed
- No coordinated effort to understand Internet and communicate strategy to customers
- [Netscape & Verity press release attached]
- Summary
Thanks to Laurie Tratt and @jackinlondon for comments/corrections/discussion.