Explorations

Capistrano for Rails

Clearly, Capistrano is a must if you're doing Rails development.  In a nutshell, it gives you painless deployment of your application.  Schema upgrades, code updates, server restarts, the whole thing.  Across all the machines in your server farm.  It's the best deployment tool I've ever seen at any price, and it's free.  Thanks and hats off to the developers!

Microsoft Still Doesn't Get It

Bill Gates is saying that Microsoft is going to start shipping IE every 9-12 months, and that IE 7 is far overdue.  I'm sure many in the business press will take this as very good news indeed, proof that Microsoft is moving in the right direction.  And to some extent that's true.

But if you read between the lines, the problem is obvious: Microsoft still doesn't get Internet Time -- they should be shipping this thing practically monthly.  Annual releases are so 1991.  Firefox ships fairly major revs far more often than that, and they have nowhere near the resources that Microsoft has.

Speaking of Microsoft, what's the deal with live.com?  Is it a joke?  Could they really be so inept at dealing with the Web?  I mean, I know it's beta but it's really, really bad.  It looks like it was done by 4 interns.  (Lest you think I jest, here's a site that 4 interns in fact did build: https://www.copilot.com/)  It's got ugly CSS, cheesy text, and it manages to look cluttered without actually providing all that much information.  And seriously, there has to be somebody at Microsoft who can generate good-looking thumbnails.  I'm frankly surprised they let this site carry the Windows logo. 

Another Vista Prediction

I wrote a few days ago that I think Vista is essentially going to be a lemon.  Looks like I'm not the only one: Steven Vaughan-Nichols, a Senior Internet Editor at Ziff Davis, agrees in a recent column.  It looks like Microsoft is screwing the pooch big-time on this one.  With Apple firing on all cylinders, it looks like interesting times again in the competition for the last mile of the computing space -- the PC.

Microsoft Retreats in the Battle Against FairPlay

Microsoft has announced that they are stopping development of Windows Media Player on the Mac platform.  The explanation they've given is absurd on its face:

"We've been so focused on Vista and we've been so focused on creating an incredible media experience on the PC," Harader said Friday. "We just did not have the resources to do a good job on (the Mac version)."

A longtime Microsoft-follower could reasonably observe that not having the resources to do a good job on something has never stopped Microsoft from shipping software before.  So this is almost certainly not about quality.  What, then, is behind this decision?  The answer is simple: Microsoft has realized that it is engaged in another platform war that it's losing, and it responded in typical Microsoft fashion.

The platform battle in question is the Windows Media DRM technology versus the Apple FairPlay technology (Google also recently began shipping its own DRM format to go with Google Video.)  It is not in Microsoft's corporate DNA to cooperate, so when faced with companies that have the temerity to produce software, its default response is to circle the wagons and go it alone.  Hence, no Windows Media on the Mac -- if people want to play Windows Media files, they can buy Windows.  (The bit about continuing support of the current version of Media Player on Macs is a distraction, since the thing that matters is support for forthcoming versions, on future OS X versions.)

The problem, this time around, is that Microsoft's technical delivery capability has begun to seriously lag its desire to own every market in software.  So while the biz people understand the strategic importance of DRM to the digital home, the tech side has atrophied to the point where it is incapable of shipping quality product in any meaningful time frame.  So while they would like to force more users onto Windows and Windows Media, what is as likely to happen is that Apple will continue to dominate media and users will defect to Mac. 

Quick FireFox Tip

This has bothered me for a while, so I figured I'd post it in hopes that it might save someone else a little frustration.  I typically use the keyboard a lot when I'm browsing -- Alt-F-C to close a tab, for instance.  I've been annoyed for a while by some websites that hijack this standard behavior for their own nefarious purposes.  A prime example is Wikipedia (see discussion here).  I think this is pretty clearly a bug in FireFox, but the FireFox developers don't seem intent on fixing it.  So here's what you can do to disable this awful behavior on your machine:

1) In the address bar, type about:config.
2) Enter "accesskey" in the filter box.
3) Double-click on ui.key.generalAccessKey and enter 0 in the popup.
4) Restart FireFox.

That's it, and you can have your browser back!

Better Flickr Uploading (or The Internet is the Platform)

Recently, the Flickr Uploadr tool for Windows stopped working on my computer.  The Flickr forums were no help, and I didn't want to spend all day uploading the 4 or 5 pictures in my set.  So I searched around a little and found an alternative: the Flickr Importr.  This tool, apparently by an independent developer, let me upload my photos in no time.  Even better, it allowed me to specify the tags and visibility for each picture.  I highly recommend it.

The interesting thing about the Importr is that it's but one of a galaxy of third-party tools built on the published Flickr APIs.  There's even a site devoted to rating some of the best at The Great Flickr Tools Collection.  And all this for a company that's fairly new and not nearly the leader in its space.  Years ago, tools like this were built to new Microsoft APIs, but it seems that more and more the innovation is occurring "in" the network.  The rise of Web mash-ups is well-documented, but that's more about combining existing Web tools than about creating new value around a single site.  On the other hand, applications like those available for Flickr, Google, and salesforce.com are built specifically to inhabit the firmament created by a single vendor.  And that I think makes them different than mash-ups.  These captive developers are strengthening the ecosystems around their core vendors.  All successful platform vendors do this, but it's interesting to see how fast it happens with published Web APIs.

We're still at a phase where developers will build to a Web API because it's available and they can build a something neat in a weekend.  But over the long term, the ability to rapidly promulgate a platform among the development community will have huge implications.  Salesforce.com really understands this, and is creating incentives for ISVs to build on their platform.  One of the most important is that it will speed the platform adoption cycle and make it much harder for a platform to remain dominance.  Intertwined with this is the idea that using a Web API allows developers to operate at a significantly higher level of abstraction than even the most advanced toolkits.  I'll write more about this topic in coming weeks.

Gates Asleep at the Switch?

There's been a lot of reporting on the recent set of Microsoft memos indicating that the company is refocusing (again!) on Web services.  It would seem at some level that they understand that the Web is becoming the platform of choice for software developers.  (Ironically, this is largely because they spent the better part of a decade making it almost impossible for ISVs to make any money on desktop applications.)  But now they're deciding that they want a piece of this whole "Web" thing in a bad way. 

So they're floating the idea of hosting all of their applications and making them available for rent.  How will all this hold up in reality?

First, they are clearly on the right track.  Most software will be a service, period.  The initial wave of Hotmails and eBays led to the second wave of Flickrs and Bloggers, along with the developer-focused Google/Amazon/Flickr APIs.  After all, if you don't have to put your users through the DLL hell that is desktop software, why would you?  The Internet has even warped user expectations regarding desktop software: people complain that iTunes doesn't auto-update itself, instead requiring a new install.  So Microsoft is belatedly getting there.

The problem is that they don't understand all the reasons why developers are moving to the Internet as a platform.  Some of the most common:

• I can ship my software every day.
• I can use whatever technologies I want on the back-end.
• It's free. 

Microsoft is bad at all three of these areas.  It's hard to make a case for Visual Studio at $200 when Ruby on Rails is free and arguably better for my particular job.  Microsoft's ecosystem is hostile to open source, one of the best accelerants for software teams.  And Microsoft has grown lethargic and is utterly incapable of shipping product quickly.  At least Ray Ozzie understands the need to be nimble:

"Developers needing tools and libraries to do their work just search the Internet, download, develop and integrate, deploy, refine," Ozzie wrote. "Speed, simplicity and loose coupling are paramount."

So what does this mean for Microsoft's Web services initiative?  Let's call it MS .NET 3.0 to indicate that they've supposedly gotten the religion before, but that they still haven't shipped a major update to their browser in a year or so.  .NET 3.0 will likely go the following route: they'll take it really seriously and put together a 'kill team' to make it happen.  The Kill Team will find no clear enemy.  The Kill Team will realize that to make the initiative successful, they have to endanger the Windows/Office licensing models.  And then it's up to Ballmer & Gates: are they willing to bet the company again?  Regardless of the details or proximate causes, that decision point is coming.  Who knows whether they will succeed?

Google As Platform

Google has officially said that their mission is to help the world index its information.  And the website plus Google Desktop have been an excellent start to this.  Google Maps at first seemed to be a tangent to their core business.  At some level, it probably is.  But at a deeper level, the release of the Maps API indicates that Google is really trying to become a "real" platform.  As more software migrates off the desktop and onto the Web, the key platforms will increasingly shift away from Windows/Mac/Linux and move toward a common Web platform.

My current favorite Google Maps application is the Gmaps Pedometer.  It's great for calculating the distance of irregular running paths!

Watch how people use the Google Maps API for examples of more Web 2.0 functionality to come.

Overview of Enterprise Software Pricing

The Economist has an overview of the state of pricing in enterprise software.  It's probably the most comprehensive I've seen anywhere recently, covering most of the important bases:

• Multicore
• Open source
• Virtualization

You can read the article here (you may need a subscription).  Also note the reference to the Azul Systems Compute Appliance, which ships with up to 16 CPUs at 24 cores each, for a total of 384 processor cores in a box! 

EU Says 'No' To Software Patents

The European Parliament has voted not to accept the software patents directive it was considering.  I'm not very well educated on this matter, but my quick take is that this isn't the big news some in the Free/Open Source community make it out to be.  For example, Groklaw hated the directive.  (Most coverage I've seen conspicuously ignores the fact that most of the money made in software happens in the U.S., which has had software patents for quite some time.  And money being made is at some level a good proxy for providing real value to users.  I hear that Gentoo is good, but commercial distributions like Red Hat and SuSe are what run businesses and give Linux much of its mainstream credibility/scalability/driver support/ease-of-use.)  For starters, it seems obvious that big U.S. companies would not have been so complacent for so long if the intellectual property (IP) laws differed so much between the two continents.  This article seems to reinforce that view, specifically with the following quote:

"Patents will continue to be handled by national patent offices ... as before, which means different interpretations as to what is patentable, without any judiciary control by the European Court of Justice," said EU External Relations Commissioner Benita Ferrero-Waldner, representing the EU head office at the vote.

This sounds like American federalism, which admittedly can be exceedingly cumbersome (think of the infinity of liquor laws in the U.S.).  But by the same token, federalism still lets important work get done.  As a bonus, the flexibility given to each state allows the country (or the EU in this case) benefit from the experience of collectively having tried many different approaches.  When one approach becomes clearly the best for everyone, it can be codified at the federal (or EU) level.  But clearly software patents aren't there yet.  So software IP will still happen at the individual country level in Europe.  This sounds like the the biggest non-story of the week.