Well, that was a lot of feedback. I was actually pleased to see so much of it articulate and thoughtful. And of course no problem at all if people want to disagree with me.  I just ask that you keep things factual and respectful.

It was interesting that many people self-identified as "net thugs" - even some who had intelligent and reasonable input. I had always thought that trolls didn't know they were trolls, or were insane, or adolescent or something. If you suspect you are a "net thug", then why keep acting that way? To those who believed I would label them as such because they disagree with me: please, keep commenting - that's not a problem.

I'm going make a few comments about some themes that came up in the feedback, since replying to each one would be a little overwhelming. And once again, this is my personal opinion. Any errors of fact are mine. Obnoxious comments are mine alone. Microsoft just makes the ID card in my wallet.

First, I'll just say that I appreciate the feedback on Word itself. Believe it or not, here in the engineering team we're very humble about the products we make. We see every failing, every lame thing because we are so close to it. I've blogged before about product development, and people seem surprised to know that we are intimately familiar with what is wrong with our products. In fact, our view is skewed to the negative pretty badly - most users don't see anywhere near the same set of issues that we see. But on the other hand, I think Word is a great tool - it just could be better. So feel free to tell me what you think needs improving. I have almost certainly heard it before or experienced it myself, but each piece of feedback is another vote to help us prioritize.

Some people were disappointed that I "gave up" on the Mac so quickly. Some others interpreted my personal frustration back then as a blanket condemnation of the Mac platform today. Not at all. I actually really like Macs and the current MacOS is a worthy product. As a designer, I particularly like it because of the thought and innovation that goes into it - it is refreshing to see that. I see that in some other products too (MS and non-MS), and I always appreciate it when I see it. Windows Media Center is another one I quite like.

On a side note, some people asked about Open Office or Star Office. As a designer I find these applications rather uninteresting. Unlike the MacOS, or even WordPerfect, there is next to no originality there. [Edit: It seems, and supporters often say] their stated goal is to clone Office97, and they are so focused on that that there simply isn’t anything to learn from or appreciate [in terms of product design]. A funny anecdote one of the people at work tells is that he was in Germany at Cebit some years ago, and stopped by the StarOffice booth (this is before Sun bought them). He remarked on how similar the UI was to Office, and the rep there nodded happily. He then pointed out that in their current beta they had the floppy disk icon in the save toolbar button backwards compared to how Office did it. The rep took note, and sure enough the shipping version had this "corrected". OK, as a designer, I find that a "desert of the mind". Please, no flames on this - it is just tiresome for everyone. I understand the story is possibly different if you are a developer. I'm just stating what it is like for me when I look at those apps. If they ever start adding original things with clever designs (that aren't clones of specs I wrote myself!), they will become interesting.

Back to my Mac problems. One of the people calling themselves "Anonymous Coward" actually hit it right on: System 7.5.2. I was a fairly sophisticated Mac user, and I knew all about extensions, and extension manager, and what sorts of things could go wrong in the Mac. The fact that the Mac I had loved for years for its simplicity and appliance-like perfection (I loved MacWord 5.1a similarly) was now forcing me to have to deal with this overhead of managing its internals on a daily basis wore me down over the months. I had to rebuild my system file regularly. Many sad Macs, and system bombs, although Type 11 was notorious. I did NOT have Mac Office 4.2 on that machine at the time, although I later added it and things got no worse or better. About six months after I got my PC at home, I was still following Mac news, reading MacWeek, etc. I came across a small item that said that Apple had released some shared library in an SDK a year or two before that was the source of all these Type 11 errors, and loads of ISV developers had incorporated that code. Yikes!

There were a lot of comments that claimed to know the "true" reason about why Word succeeded or similar claims. The reality is that the marketplace is a complex system, with a lot of variables. There is no "true reason". Analyzing this system is something I find fascinating. In a footrace, the "winner" is the person who is first to cross the finish line. Do they have to beat the world record to deserve the win? No, they just have to be faster than the other guys. In the same way, a product that succeeds does not have to be perfect - it just has to be "better". That means not everyone will love it (Word has 400 million users  - if just 1% of them disliked the product that is 4 million people ready to swamp any blog with their anecdotal experiences). And for the record, I'll state right here that Word is, like any product, not perfect and is actively being improved. We can't fix everything - we can't do all we would like, it might take more than one release, but the various problems people have mentioned in the comments are known issues that we hope to address in the future. If you're interested in why products aren't perfect or don't have the feature you want, I have blogged earlier about the process of building products here and here.

What does better mean? In a marathon, one runner might have greater potential, but another runner might have better strategy - drafting off the leader, conserving energy - maybe even "psyching out" the other guys with a late unexpected burst. In running, this sort of strategy is considered part of the sport, and athletes with strategic skills are lauded for that. In business, it is also normal - you compete with what you can bring to bear. Timing is critical. The point is that any market is a fluid thing, and having a great product is only part of the equation - although it is pretty hard to win without a product that is at least close to the best if not the best.

With computer products (maybe this works in other industries too), there are windows of opportunity. If you can latch on to one, then you will succeed beyond what some might say are the merits of your product. A good example is "VHS vs. Beta". A popular fable is that Beta was a better videotape format than VHS, but VHS won anyway, so the claim is that it is not true that the best product always wins in the market. But in this case, the definition of "best" had a temporal element to it. There is a technical argument that beta recording quality is slightly better than VHS, but most consumers couldn't really appreciate the difference. What mattered much more was that in 1977-78, VHS tape was the first format to come out with 2hr recording capability. Beta could only do 1hr. VHS held this advantage for most of a year. 2hrs vs 1hr is not just twice as much - it is a movie vs. half a movie. Consumers picked the product that was "best" because it did what they wanted - let them record movies off TV. Later, Beta came out with 2hr tape, but the momentum had shifted from the early mover Sony to the followers (VHS consortium), and Sony couldn't keep up. Was that the only reason? No - Sony also didn’t license their technology, and kept their prices high. But for those who study this sort of thing, the decisive factor was the 2hr vs. 1hr thing. Sony missed their window.

In my case, I had a "tipping point", and for such a dedicated Mac guy as I was, it was heartbreaking to have my Mac just tank constantly. (yes, I know about "Save", but for the poster who made that comment - surely you would not accept that answer from me if you had a similar problem - use your head dude!). Once I had switched to a PC at home, there was no going back - I found there was a lot of software for the machine - all the latest games, whatever (and this from a zealot who argued "who needs 100000 apps - the Mac has two of everything you need"). My PC worked fine, and there was no reason to go back to the Mac, even if System 8 had resolved things (I never actually checked).

Some of the posters noted that Word was helped to success by the Office bundle. That is certainly true - that move was a truly inspired marketing decision to use our strength of having enough apps to build a "suite" - something which hadn't existed up to that point. At first it was just a bundle of three apps for the price of 1.5 apps or so. People said it was crazy - too much of a giveaway. At the time, word processors such as WordPerfect, Word and others sold for ~$500, so to sell Word, Excel and PPT for only $699 seemed illogical. But in fact what it did was make the price of getting all our apps much more affordable, and we often got $699 (discounted of course) from people who would only have bought one app before for $499. This is called "supersizing" in another industry :-). BTW, the price of Word and Office has come down nearly every release - as some people commented Microsoft led the way in making software cheap to "make it up in volume". There is another myth that Microsoft made software as a whole get more expensive - one comment actually raised that classic. That one is beyond me - no factual evidence exists to support that assertion that I am aware of. Three apps purchased separately from the market leaders were around $1500 before Office. Office is now about $379 - nearly half what it was when it came out, and that is in current dollars. If you keep Office for 3 years before an upgrade, you are spending ~$8/month for software you use most of the day to get your work done or run your business. That's three lattés, or two if you get syrups and extra shots. It is also about 4 days of cell phone usage, or cable TV. Pretty good deal in my mind.

Other commentators said they thought the Office bundle was the primary reason for Word's success, because they felt Word itself was second rate. I don’t doubt the people who hold that opinion since everything's subjective at the individual level, and there is no right or wrong about what people feel, but I have a whole bookcase full of "best word processor" awards starting from 1990 across from my office as proof that the great majority of reviewers agreed that WinWord was the product with the most right stuff at the right time.  In fact, there is a strong correlation with a product winning reviews and the start of its ascendancy in the market. It happened with WordPerfect, it happened with IE, it happened with Word. Each of these started winning reviews as "best in category", and within 2 years they were market leaders. Was that the only factor in these products' success? Certainly not. But at least the market is consistent in this respect - the products that are considered best by the "qualified" judges end up on top. (I suppose the IE comment is going to generate a lot of flames. Oh well - it is simply a fact that IE3 won 19 out of 20 major reviews in the space of a year after it came out when it was put head to head with Netscape 3 and whatever else was out there. I'll leave it at that.)

I also detected another old saw about hidden advantages or undocumented APIs that somehow made Word better than competing apps. The reality on this is so counter to the conspiracy it is astounding. The Office team barely talks to the Windows team. If anything, there is more mutual contempt than cooperation, although we constantly try to make that better. And, what are “undocumented APIs" really? Are they APIs like “ExReallyFrickingAwesomeTextOut()" that somehow make our applications have features and usable designs that customers want? Actually, Michael Cowpland (CEO of Corel at the time) said it best in an interview in LinuxToday a few years ago. I don't know how I saw this, but I kept it because I was astounded that he actually shared it given that it was in his interest to maintain the myth:

Dwight: There was no issue that Microsoft Word would be able to run better on Windows than WordPerfect because you weren't getting information?

Michael: No. We know that for a fact because with our JBridge technology we had to X-ray the OLE APIs and I actually asked our chief engineer doing that what about all these undocumented APIs, the ones they're supposed to use and there truly are some undocumented APIs but it's ones you wouldn't want to use anyway. They are just for internal communication. I had him print them all out and we could see they weren't sinister ones. They were just there because they needed internal communication and there wouldn't be any point in exposing them because you wouldn't want to use them anyway.... “

To me this is a fascinating example where doing the actual fact checking is hard, so people never bother. But in this case, they had to, and they couldn't believe that the myth was untrue. Makes you wonder about all these myths - if they were true, there'd be hard evidence of some kind - like a dev coming up with the supercallifragillistic undocumented APIs. But there isn't, so...well. Now, Michael is a character in his own right - if you're Canadian like me you know that his ego was having trouble fitting in the Ottawa valley for a while there. Trophy wife, Lamborghini, mansions, fiscal impropriety *ahem*... Well, events seem to have addressed that.

Another set of comments were on file formats. This is one of my favorite myths, in the sense that so much is written about it, speculated, then repeated endlessly as "fact", and there are only a handful of people who know the real story. The real story is also so boring that it has no chance of ever overtaking the "meme" that carries the conspiracy theories. It is sort of like the Bermuda Triangle. Many people actually believe that an unnatural number of ships/planes are lost there. There is also a lot of documentation backing up all sorts of stories, with official sounding sources such as air-traffic control transcripts, etc. If you have ever seen one of those debunking documentaries, you know what the real deal is. Everything is just quoting other stories that quote other stories - except for one or two original documents - and these are pure fabrications. The air-traffic controllers never said those things, the transcripts don't exist, and the actual humans in the control tower swear none of that transpired. But yet a whole culture of people believe it because it is printed and documented with citations and everything. If you repeat something often enough, it must be true.

So, file formats. Here's the deal. In the 80s and early 90s, every time (nearly) that a new version of a product came out (1-2-3, WordPerfect, etc), the standard deal was that the new version had a new file format. This was a no-brainer and was considered normal and acceptable by the market. The product had new features that the old one didn’t understand, and you ran these things on standalone machines that had no network, so as long as the new version could read your old files, you were golden. You almost never had to send a file to someone else electronically - you printed it. Innovating meant a new file format.

So, this was the tradition, and generally applications followed it. Word6 had a new file format compared to Word2, as was normal. Then Word95 came out - it was a 32-bit port of Word6 with only a few new features (although it had my all-time favorite - background red-squiggle spelling), and none of these affected the format, so it wasn't changed. Word97 was started in 1994 at the same time as Word95, and almost the first thing that was done was the routine change of file format to accommodate some of the big plans we had (Unicode support being a huge one). Something really big happened in 1995 though - the internet, email, and Windows95. Suddenly everyone was getting a computer to access the internet, do email, and/or to experience the "wonder" that was Win95, since that had been such a big deal. PCs were becoming mainstream and were spreading everywhere. Corporations were deploying them in huge numbers. Another development was accelerating - LANs, and WANs - so electronic copies of documents could be shared inside companies much more easily.

Around the beginning of 1996, well past "code complete" for Office97, we started to realize that the world had changed. Word6 had been sold into a market of about 10 million consisting of pretty techy users with few interconnections except via floppy disk ("sneakernet"), and it was the "challenger" product, so the installed base or older versions was small. Word95 had been unnoticed since it had the same file format, but it was widely adopted - not as much as Word 6 though. Word97 was going into a market of about 50 million, and Word was now "the standard". We quickly tried to do something about the impending problem, but it was really too late. We had to ship without the ability to save the old binary format of Word6 (there is only one binary save path in Word and it is quite baked in), and of course we couldn't go back to the old format since it would mean removing most of the improvements: all the new graphics, international support, etc, not to mention a huge delay. We started a crash project to build a "downrev save" converter to the old binary format using Word95 as a base, but that wasn't ready at launch. Thus was born a legion of conspiracy theories about our "true" motivation for changing the file format.

The reality of that was that customers were pretty dissatisfied, and wouldn’t buy the new version. Sales stalled at first, and we made a rule that the next versions of Office had to save in a format that was compatible with 97. We could still add new features, but whatever they were, they had to fit in the old format. This is why Word97, Word2000, Word2002, and Word2003 all use the same binary format. Fortunately we had those last few months to add some bits to the 97 format that made it possible to add things in the future that the old versions of Word would ignore politely, but sticking with the same format for the last 8 years (Word97 shipped in 1996) has put a significant crimp in our style. There is a corresponding claim that circulates the net that says we change the format "every release". Since we bend over backwards not to do this, that one always makes me chuckle.

Another one I like is the use of the word "obfuscation" when some people on the net (usually the "thugs", or "anonymous cowards" as they describe themselves) describe our behavior with the file formats. The dictionary on my desk (actually the Encarta dictionary built into Word2003) says obfuscation means "to make something obscure or unclear, especially by making it unnecessarily complicated". This implies some kind of intent with the word "unnecessarily". In fact, the file format for Word is simply the most convenient way to save files for us. That's it. It appears complex to others because it is highly specific to Word, and it has undergone all sorts of contortions to try to remain compatible over the last four versions. We also have documentation for it (I can find some older versions of it on the net published without permission - I bet you can too). We make the documentation available to partners, governments, anti-virus vendors, etc. Some people ask why we don't make it public, often in a tone that implies we are somehow required to do this ethically. We don't do that because it is our intellectual property. People who want to work with us can get it by contacting us; people who want to compete with us need to work harder. That's business. We might change our minds about that if it seems that making the format public would be of most benefit, but really it is our prerogative.

By contrast, Word has supported interchange formats for years. RTF was created as a way to exchange formatted text between applications, since before that there was no widely accepted way to transfer anything other than plain text. Then HTML came along, and although at first it was so basic it couldn’t really handle word processing formatting, it got better so that now it sort of handles it, but not completely.

HTML is another interesting story. We take a lots of grief over our HTML support in Word. But it does what it was designed to do. The problem is people don't realize what we intended it for. Word HTML was designed in 1997 (mainly) as a replacement file format for Word. As such, it was not designed to be created by any app but Word. The idea behind it was that you could create intranets where people save documents as HTML to web servers directly (not going through an admin), just as you would to a file share. They would be browsable as web pages but still editable as Word docs. Most people still don’t know you can open files off a web server directly in Word, edit them and save - but it is pretty cool. We built our product specs that way for several releases. We had a problem though - in order to make the experience seamless, we had to be able to round-trip everything Word normally saves in its binary format through HTML. This includes little details like which words have already been spell-checked and other niceties we store to provide a better user experience while editing. But HTML had no concept of many things that Word did. So we used the extension mechanisms of HTML to enable us to save things in HTML that it couldn't normally handle. We followed the rules as we did this (bugs notwithstanding), but the resulting HTML was not like anything HTML coders had seen. They were used to minimalist, human-readable HTML. Being human-readable was not a goal - after all, most people don’t care what HTML looks like when they look at a web page. Thus started another huge myth, that we were out to somehow "co-opt" HTML with Word, when in fact we were just trying to make things work in the confines of a very limiting format. In a later release we added the ability to remove much of the "round-trip" info for the people who wanted to use the HTML we made in places we didn’t design it for (this is called "filtered HTML").

XML is also a fun one. We designed our XML support to be what we had long hoped HTML would give us - a great interchange format that was clean and designed the way we needed it to be to be optimal for Word. Word2003 has a new optional file format called "WordProcessingML" (WordML for short). This format was designed to be manipulated by external processes, such as server apps assembling documents from fragments. It is also a full - round-trip format like HTML was, but it doesn't have all the contortions we had to go through to make HTML work. We also wanted to support schemas that a customer might define for their data. So we added support for XSD, which is a standard that the W3C defines for describing data. If you have the Pro (business) version of Word 2003, you can markup a document with a schema of your own design, and we are able to export this in exactly your data format. This enables all sorts of interesting interactive "applications" you can build on top of Word, from invoice tools, order management, assisted authoring, work flow, etc. We built a little tool internally to help write our specs for the next release of Office, and it uses WordML as the document format, with another schema on top designed to carry spec-related markup in the document, and aggregate the spec info on a SharePoint server (e.g. you want to see all the open issues in all your specs - this is now quite possible with some server-side script pulling data out of the WordML docs on the server).

We had some funny moments last year when we announced the XML support, and that it was going to be fully documented. To us this was natural, since why on earth would we do all this work only to not document it, or make it hard for people to use? But when the announcement came, the conspiracy people on the net reacted in shock. Since they had built-up a complete alter-ego Microsoft that had bad intentions for everything it did, this caught them by surprise, since it didn’t fit their world view. This was when I really got to know Slashdot. For kicks we would read what people said there, as they struggled to twist reality to fit their view, and speculated on what (me and my colleague's) intentions were. It was surreal. The moderator scores were especially humorous. About one poster in 100 would actually get it right, but their posts would be rated something like a "2", and completely baseless conjecture was rated a 5. People would sometimes post real samples of the XML Word produced to try to bring some clarity to the discussion, but that was rated low of course, and always killed the thread - it seemed it wasn't fun to deal with the fact that the XML was real, standards compliant, and useful. So a new thread started, with the same rants repeated. But not to worry, eventually a way was found to incorporate our XML work into the conspiracy world, so all has returned to normal.

Well, it’s getting late, and baby seems to be nodding off again finally, so I am going to get some shut-eye. I'm sorry if I offended anyone - I just thought it might be interesting to hear a side of things that you probably don't get to hear. Next post I'll take up topics that people are interested in, and try not to get distracted by the “thugs”.