Tuesday, July 01, 2003

Bad Blows The technology world has taken a couple of bad blows recently in court. The first happened a couple of weeks ago, when SCO decided to attack Linux. The second happened last week, when a U.S. Circuit Court of Appeals ruled that Microsoft does not have to bundle Java in the operating system. These two unrelated events may seem opposite in nature, the first being a small company winning a victory over a large corporation (IBM), and the second being the largest of companies winning a victory over a smaller competitor. However, these two issues actually go the same way - against the best interests of technology. I believe that technology has reached a point where it is very difficult to innovate. Innovation comes from one of three sources: the big companies, small startups, or the standards/open source community. When big companies innovate, it tends to be clouded by their own technology vision, which biases it, and therefore makes it less likely to represent true innovation. When small startups innovate, they are held to very high revenue standards in this market, while potential customers are less willing to buy from their risk profile. This means that very few survive, and those that start to survive are usually bought up by the big companies (see above). The last category is the true hope for technology innovation, in my opinion. Sure, there are politics and biases in the open communities, as well, but they are balanced by the nature of the individuals involved. So if we are going to innovate in a economic environment like this, the most likely source of innovation will be the standards and open source groups. Both of these court events I mentioned are tied to this, and both of them hurt open source. In the first, you have a situation where an open source platform has actually become mainstream. One of the big companies (IBM) has adopted it and helped to move it forward. Then along comes charges of copyright infringement. This is the worst thing that can happen to open source. Big companies are already worried about hidden issues when they look at open source. The potential for copyright issues just adds fuel to the fire. In the second case, the most pervasive desktop software in the world is allowed to eliminate Java from its platform. I have less of an issue with this one philosophically, but it is going to hurt Java. Corporations, already concerned about Java's weakened presence on the desktop, won't even consider client Java products now. So this trend is not good for technology innovation... 10:24:26 AM

	Monday, June 23, 2003

Operator Friendly I've been doing a lot of work with two IBM products over the past couple of weeks, WebSphere Portal Server (WPS) and WebSphere Studio Application Developer (WSAD), and it strikes me how very different these products are in terms of their ease of use. It amazes me that they actually come from the same company. Then again, maybe I know the source of their divergence... First, let me explain what I mean. WPS took me two full days just to install, including all of the pre and co-requisites. The install process was fraught with pitfalls, and I had to uninstall and reinstall several times, manually updating my Windows registry, searching out obscure fixpacks on IBM's FTP site, and sorting through four different documents that explained the installation steps (different steps in each, by the way, and none of them completely correct). Once I was up and running, I ran into further configuration issues for the following two days. What is truly astounding is that this is, by far, the leading portal server on the market. In all fairness, once it is up and running, it is relatively easy to use, but getting it running is a horrendous experience. I have heard that IBM's own technical reps do not install it, but rather send their laptops to support to have it done. I wish I could have. Now, installing WSAD was like a dream. The product installs its own application server environment, sets up its libraries, and allows you to include a bunch of detailed examples, each very self-contained. Never did I have to go elsewhere to get anything. It was the perfect installation experience. On top of it, the product is extremely operator-friendly. It allows you to view your application from a variety of perspectives, and switch perspectives instantly. It is simple to import libraries, move objects between projects, and even refactor changed code - without ever worrying about classpaths or compilers.  When building web applications, deployment descriptor files and configuration files are automatically generated, and can be viewed and modified through intuitive, structured editors. Best of all, the compiling and packaging is done for you as work through your project, allowing you to focus on writing code. I would have to say that this is the best software product I have ever seen from IBM. (That isn't necessarily saying a lot, because I believe that very few excellent software products have ever come out of IBM. Most of them are very utilitarian - they work well in terms of reliability and speed, but are very difficult to work with. But this product is really good.) So how is it that two products come out of the same development organization and yet they have completely different levels of operator-friendliness? I think there are two reasons for this. First of all, much of WSAD was not created by IBM. It comes from the Eclipse project, which is open source. I think this model works really well for IBM because the open source community is particularly good at the things that IBM is not that good at - primarily, making things easy to use. IBM, conversely, is good at making things that work well and that are stable. The combination produces very nice results. WPS was produced entirely inside of IBM, and you can tell. It works well and is very reliable, but getting it running is no small task. The second reason is that WPS is targeted to a different audience, at least in the minds of IBM. WSAD is made for developers, while WPS is made for IT departments. (I think that IBM is missing the point a bit, because I believe developers ultimately influence these decisions more that they think - but that is a point for another article.) I think that IBM does not put a high priority on operator ease of use because they know that the guys who have to swim through it are not the buyers. That is a bit pessimistic, but I believe it is true. To take my pessimism a step further, I also believe that IBM is intentionally not fixing the problem because it is a source of services revenue for them. What happens is they sell at an executive level - either on the business or on the IT side. They are able to sell there because they are IBM - you are buying the name - nobody ever got fired for choosing IBM (you get the point). The buyer forces WPS on the implementation folks, who quickly determine that this is too hard to use and that they are going to need some help. IBM is happy to step in with gobs of services (read: more revenue). This is what always happens with "Enterprise" products. So maybe IBM will never change when it comes to those high-end product sets. Why make them easier to operate if it only serves to cut revenue from your services? At the opposite extreme is Microsoft. I believe this is precisely why M$FT has continuously beaten the pants off of IBM on anything that the user has to touch. Perhaps as portals get closer to the user, and Microsoft figures out that it needs a real portal play, IBM will be forced to change. If they do, I think they should use the model they used with WSAD, because it really produced a nice result, and for the first time IBM has produced a user-oriented product (in this case the user is the developer) that is competitive with Microsoft in terms of operator-friendliness. 11:39:14 AM

	Friday, June 20, 2003

'78 Ford Grenada I've been reminiscing recently about my old high school car. It was a beast,but it had its positives. Sure, it was ugly and uncomfortable, and yes, it would often stop working. But when it did stop working, I could fix it. I changed the clutch and the brakes, always changed my own oil and plugs, and I knew how to get it to start when it didn't want to. It wasn't that I was some sort of genius, it was just that it was all there - accessible, assembled with standard tools, and purely mechanical. I open up the hood on my car today and I am completely lost. Electronic ignitions, fuel injectors, computers - even some of the bolts and screws require a specialized tool. If my car breaks down today, I'm screwed. Somewhere along the line, we lost that connection between man and machine where we completely understand how it works. Where everything used to be made up of small discreet components whose function and mechanism that were understandable to the layman, now we open our hoods to a black box that very few people can understand. True, the black box exposes way more of its diagnostics to us in the form of intelligent messages and warnings (my Grenada had one dashboard light labelled  "engine"). But when those systems fail, or even when they tell us something that requires action, we are forced to go to an expert. So are we better off? Well, I have to admit that I would rather drive my Volvo than that old Ford Grenada. It is much more comfortable, has many more amenities, and ultimately, forces me to open the hood much less often. But if I were stuck in a world without mechanics, I would take the Grenada. I think the same argument can be applied to the Unix versus Windows discussion. Windows is more comfortable, has more amenities, and forces the user to get beneath the hood less often. Unix, like my Grenada, is completely accessible. It can be ugly, and may not have all the bells and whistles, but if I need to fix it, I can. Now, I know some people will disagree with me. There are some who are much more comfortable with Unix, just as there are some who still prefer the old cars, and for much the same reasons. But for the most part, the mainstream would agree with my statement. So why is it that we can't have the best of both worlds. You could make the case that Linux is heading in that direction, although I think Mac OSX is closer to what I am looking for. In any case, don't be surprised if you see new cars coming out in the future that return some level of control to the home mechanic. Then again, most in our society would much rather sit in an auto shop lobby than wait on the end of a customer support line... 10:45:16 AM

	Friday, March 28, 2003

Massive Virus Found in Water Cooling Tower Scientists have discovered what they think to be the largest virus in the world, larger than some bacteria. It was found in a water-cooling tower in Bradford, England, feeding on Amoebas. Is it just me, or does this scare the hell out of anyone else. All I can imagine is this tower incubating mankinds' death. Now, they don't know if this virus even has a physiological effect, but they do think it can infect humans. I am inclined to believe the work of Dr. Frank Ryan, who claims that the end of the human race will come at the hands of a virus. The mutation process is just too quick for our internal antibody engines to keep up. I don't even want to get into antibiotic abuses and malicious bioengineering efforts. We can worry all day long about some big meteor hitting Earth, catastrophic global warming, or any number of other horrors from beyond, but I'd put my money on the little guys to take us out. Sources: NewScientist.com news service 19:00 27 March 03, Virus X (Little, Brown) 10:20:03 AM

	Thursday, March 27, 2003

Bad Technology Options I was talking with some of my friends today about the lack of good technology companies out there while lunching at the Rainforest Cafe. It is strange that there appears to be a lull right now. I think it has to do with a few converging factors: standards efforts, bad economy, and industry consolidation. Most of the attractive innovation is happening in the standards and open source communities right now. With the bad economy, companies are slow to adopt new technologies. The big boys, like IBM and BEA, simply wait for the standards community to define the base requirements for a new technology, then they build a product to it. This is slower than rushing into the market with something, but the slow corporate adoption is allowing them to do it. Meanwhile, startups blaze the trail with products that end up failing, offering IBM a reference model in the worst case and potentially a head start through low-cost acquisition in the best. Of course, then there is M$FT, which completely ignores the standards - or more accurately adopts the standards and then modifies them slightly so that they become proprietary. 4:31:07 PM

	Wednesday, March 26, 2003

The Dark Side I've been thinking about crossing over to the dark side. I have an opportunity to go work for a large financial institution. The prospects of stability and order are appealing to me right now, given the chaos from which I just left. The interesting thing about software right now is that it is really getting boring. There are very few small players. Most good ideas get either subsumed into standards, or crushed by the big players, and most enterprises are increasingly skittish about investing time and dollars in a small company's technology. From what I am seeing, everything will continue to be dominated by M$FT and IBM. Everyone else is just fooling themselves. The best a small company can hope for these days is simply to get purchased by a larger player, and the multipliers suck these days. Don't get me wrong, there are good companies out there with good technology, and some may actually turn themselves into viable businesses, but the issue is that today's pricing pressures do not allow startup software companies to turn themselves into real businesses. Let's take a look at what I mean. Average deal sizes for software have dropped, even for the big boys, down to about 20% of what they were. Most smaller companies are feeling this pressure. In addition, to get into deals, these companies also are often forced to piggyback off of a major players' deal, further squeezing down the deal size. At the same time, sales cycles are not reducing. It still takes at least 3 months to sell any software, unless it is bought through an online store, and three months is the absolute minimum. The median is in the 6-9 months range. So these companies hire a bunch of salespeople (who are not cheap, by the way). Let's say the average salesperson gets 2 $100K deals a quarter, but takes six months to get productive. That means that these companies are getting four deals per saleman in the first year. The costs of running the company are still high, meaning that burn rates are still running at $300-$700K per month. There is no way that a company can make it under this model. Your options are to hire more salespeople (which fails because they begin eating into each other), or to drop your prices and use a different model. This latter option seems to be the most effective. Go for mass distribution at a lower price tag. Only certain types of products can manage this, but it is a good strategy. Distribution can be channel-based (A big partner sells you as part of a bigger sale), or distributor-based (online store, etc.). There are cases where companies can get the best of both worlds, but this is extremely rare. I know of several software companies that have deals with larger players that sell their technology, while still retaining large $100K+ deal sizes. This requires that the companies are structured around a sell-through model. So anyway, my bottom line is that until things change, most of these small startups will fail. They need to find distribution models that allow them to overcome the software sales dilemma. This can be accomplished using either mass distribution (very low price tag), or channel distribution (higher price tag, but harder to get the relationship). This is why I am considering the dark side. 11:03:16 AM

	Monday, March 24, 2003

Where the Wild Things Are... I recently lost my job when my company closed its U.S. operations. Sad and strange from the inside, particularly given the remarkably positive results we were seeing. Funding, sadly, is a bitch right now. So I have had a great deal of time to watch FOXNews coverage of the war. I am very impressed with our military leadership, particularly when they give open press briefings to hostile foreign press. It amazes me how much so many people in the world hate the U.S. We have our faults, and we've certainly made our share of mistakes, but we are good people with good intentions at the root of it all, despite the fact that other nations see us as being driven purely for economic gain. 3:33:11 PM