Mike Coyle’s Weblog

…it is from kicking the shins of the giants around them.

I’ve read a couple of blog entries from folks coming back to Python after jumping on the Rails bandwagon, citing the temperament of the community as one of the reasons.

I never thought much of it, until I read the DHH’s response to Joel Spolsky’s Language Wars article, and the comments that follow. Sure, there are a number of comments that are constructive and forward looking. But you have to look for them. Carefully.

I think that Joel Spolsky has done a lot for the software development community, and was not paid the respect that he deserves in DHH’s response and a multitude of reader comments that followed.

I’ll be honest, I’ve considered jumping on the Rails bandwagon. Several times. The thing that has held me back is the notion that Ruby/RoR hasn’t really gone through its adolescence yet.

If you look at the folks who are doing great things with RoR, lots of them (37Signals, the Pragmatic Programmer guys, etc.) would be writing great software with _any_ tools that they were using. It’s like the fact that Chet Atkins could make a toy guitar sound great.

Maybe it’s my ignorance of the inner workings of the RoR community, but I get the sense that RoR hasn’t had its “holy $#%&” moment, when the hard lessons are learned by average-to-good corporate IT developers, and the sound of pagers at 3AM drowns out the cheers of the language evangelists.

I think that Microsoft had its when they started playing Paper-Scissors-Rock with DCOM and Java. I think that J2EE had its when the Microsoft Pet Shop spanked the canonical best practices of heavy EJB development. Python had its when the momentum of RoR highlighted the fact that there were 80 web frameworks that each did 80% of what people need, and no knee-jerk answer for which one to use for any given application.

But the Rails community? Not so much. I don’t think that the RoR honeymoon is over just yet. Someday it will, and the hard lessons that come will forge the legitimate counter-arguments to Joel’s points.

Yes, I am referring to Rocketboom, which debuted on TiVo this week. It shows up on the “Now Playing” list just like any other show. Here, you have (relatively) low-budget video content showing up alongside mainstream TV programming as a first-class peer.

This is clearly a testament to RSS and its descendants (blogging, podcasting, whatever-we-end-up-calling-podcasts-with-video, etc.), and a sign of things to come for mainstream media in general.

If you have TiVo (and a broadband connection), here is more information on the Rocketboom Video Download Trial.

Python.org has a number of very interesting Success Stories of how various organizations are using Python to solve real world problems. With all of the focus on web development frameworks and such, it’s interesting to see the diverse set of problems that Python is addressing every day.

In our new case study titled Botonomy Uses Python to Create ProjectPipe.com for Web-based Project Management, I give a high-level overview of both the “why” and the “how” behind ProjectPipe.com, our hosted project management solution.

The article talks about the problem that we set out to solve, and walks through one of the more detailed “behind the scenes” overviews of ProjectPipe that we’ve published to date.

Comedy Central’s annual celebrity roast was on this last Sunday night. This year’s celebrity roastee was William Shatner.

Why is William Shatner famous? Let’s just say that if you and your buddy with the pointed ears ever decide to book a discounted trip back via to 1983 to watch an action-packed police drama where the show’s protagonist sings a really creepy rendition of Elton John’s “Rocket Man”, then Shatner’s your man.

If you missed it, well let’s just say that all you Star Trek fans should set your phasers to “Incessant Reruns”. Truth be told, they might as well rename the network “Roast Central” for the 11 months following the initial broadcast of the roast each year.

This same weekend I stumbled upon the Web 2.0 Logo Creator, which I used for the Shatner logo above. It dawned on me that both William Shatner and Web 2.0 share the following characteristics:

1. Both are immediately identifiable, almost iconic in nature

   As any Star Trek conventioneer can tell you (sometimes even in Klingon, nevertheless), Shatner is larger-than-life to the Trekkies out there. I was never a big fan of the show, but to its fanbase Star Trek provided an entertainment experience that was far more enriching than that offered by its late-sixties television contemporaries.

   Likewise, Web 2.0 applications are very distinctive in their feature set, branding, and approach. To many, Web 2.0 also provides an experience that is far richer than that offered by (cough) Web 1.0 applications.

2. Both served a necessary role in the evolution of their respective ecosystems

   Shatner’s role in Star Trek helped bring Science Fiction to mainstream TV and cinema audiences.

   Nowadays, many Web 2.0 applications are causing the IT world to rethink the very nature of thin-client applications and the possibilities thereof, as Web 2.0 has mainstreamed a variety of technical and design considerations.

   At the risk of making Newton roll over in his grave, the “killer app” of tomorrow will see further by standing on the sans serif shoulders of beta-quality giants.

3. Both have managed to become caricatures of themselves, while they still “keep on keeping on”

   Sunday’s roast made it crystal clear that Bill Shatner has his self-deprecation switch stuck on the “ON” position. Yet his acting career continues to flourish in shows such as Boston Legal.

   Similarly, there have been a number of sites that make fun of Web 2.0 conventions, yet nearly all of the interesting web apps as of late embody many of the principles, technologies, and design considerations that are often subject to parody.

So, Messrs. Shatner and Web-two-point-oh, keep in mind that in the IT world, like the Friars, we only roast the ones we love.

TextMate won the 2006 Apple Design Award for Best Mac OS X Developer Tool. Congrats to Allan and everyone who has contributed to the project.

I was a Vim guy for many years, and recently made the jump to TextMate. I’ve had a few false starts in the past, going back to what’s familiar, but I’ve completed the transition in less than 210 days. (Applause).

There are a couple reasons why I switched. Most important was my belief that while Vim is great at slicing/dicing text, TextMate was built for slicing/dicing higher-level concepts and structures expressed in text. I convinced myself that the context-awareness and snippet-fu that TextMate brings to the table would lead me to be more productive in the long run.

TextMate brings to the commoditized world of text editing what OS X has brought to the commoditized world of operating systems. I think that the tastes and preferences that lead a developer to buy into and evangelize the Mac platform are the same ones that lead that same developer to buy and evangelize TextMate.

A quick analogy: If you look at text editors as operating systems, then TextMate is the OS X of Text OSes. Much like OS X takes the power of Unix and makes it lickably good looking and easy to use, TextMate has taken shell script pipelining and built the necessary abstractions required to profoundly flatten the learning curve.

But, you may rightly say, Emacs has had rich programmability for years. Any Emacs (or Vim, for that matter) guru can duplicate many of the functions that TextMate provides. Not only that, they can take the list of reasons why their editor is arguably better, have the items psychoanalyzed, and emailed to you in an encrypted file, all without leaving the confines of their Text OS, er, text editor.

My response? “Exactly.”

Any Emacs guru can and will do those things, and they will be very happy and productive in the process. Vim, Emacs, and TextMate are all great editors. It’s no longer a matter of what’s the best editor, but rather, what is the best editor for you?

I think that Emacs is the Linux of the Text OS world, whereas TextMate is the OS X of the Text OS world.

Specifically, the things that make users pick TextMate over Emacs are precisely the types of things that make a user pick OS X over Ubuntu Linux as their operating system of choice.

So once a developer makes the deliberate move to OS X, making the subsequent jump to being a TextMate customer is both easy to do and consistent with their underlying value system.

So that’s why Allan has sold scads of licenses. He’s built a great product that strongly resonates with many of the tastes and preferences of his entire target market.

Or…maybe it’s just that the Rails and TurboGears screencasts were flat-out cool. What do I know?

There was a bunch of news coming out of the Apple Worldwide Developer Conference (WWDC), but in my opinion, the story with the lowest “press-coverage-to-significance” ratio is around Apple’s iCal Server. It turns out that the server itself is written in Python using the Twisted networking framework.

Yesterday, it was also announced that Apple officially joined CalConnect, the calendaring and scheduling consortium. The iCal Server supports CalDAV, the up-and-coming standard for standards-based calendar and scheduling over commodity web infrastructure. For more info on CalDAV, check out this CalDAV article on NewsForge.
I find this interesting from three perspectives:

1. SMB Market Expansion: Apple’s movement into the calendaring space will provide another viable alternative to Exchange, especially for new businesses. Before WWDC, the rumor mill was buzzing about Apple getting into the Voice-over-IP (VOIP) business. I think that this is the classic Apple play: Take a useful technology that works, package it in an elegant interface, make it easy to use, and then cast the problem so that people say “Why didn’t it always work this way”. An easy to use Calendaring and VOIP solution from Apple would be a game-changing moment.
2. Ruby vs. Python: It puts some context around the announcement that Ruby on Rails will ship with Leopard (OS X 10.5). The bundling of Rails is great news for everyone involved. The Rails community is very much pro-Mac (all of the core commiters, and the vast majority of Rails developers work on Macs (much to the delight of Allan Odgaard, developer of the phenomenal TextMate text editor).

   However, the fact that Apple is building new, high-profile applications using Python should indicate that Apple is not likely to be a one-trick pony when it comes to dynamic languages.

3. Additional validation of Twisted: The fact that Apple is using Twisted as the foundation for something as strategically important as Calendaring should reinforce the well-deserved image and reputation of the Twisted team and the software that they produce.  In time, this is certain to introduce Twisted to the greater IT industry, and make Twisted-based applications much more interesting to the broader market.

I wouldn’t fret too much regarding the news that Monopoly will be replacing cash with Visa-designed debit cards.

Now, I don’t want to go off on a rant here, but I think that the Parker Brothers don’t know their thimble from a hole in the ground if they think that replacing Monopoly money with debit cards is a good idea. Besides, what will Canadians use in their vending machines?

Our entertainment overlords at Hasbro are just setting us up for inflated prices for “Monopoly Classic”, when it makes its glorious return. They are already counting on us waxing nostalgic for getting paid in $1s and $5s when little Bobby saunters into the beautiful hotel you built at Casa Del Ventnor

Tedium, we hardly knew ye.

Besides, I don’t understand the parents that would buy a cashless Monopoly game to begin with. I mean, Monopoly is a little slice of traditional, old-time Americana. Just last night, we sat around our well-worn Monopoly game for hours. You know, the Bass Fishing Lakes Edition.

I hear that they also have one related to Las Vegas. Or maybe it was Atlantic City.

I forget.

These days, there is a special edition of the game for every movie, city, sports team, hobby, special interest, and topical skin condition. Obviously, brand integrity is not important to either of the Parker Brothers. Seriously, you could drop a half-packet of Splenda into a Big Gulp-sized cistern of coffee, and still have less dilution than the Monopoly brand. What gives, muchachos? I mean Peter Parker is obviously busy at night, but the other one, he has no excuse.

I was at the toy store the other day trying to buy Monopoly for my nephew. I had to ask the kid at the front desk where they kept the Monopoly edition of Monopoly. His answer: 1987.

This latest move is a pure unadulterated marketing move by Visa. I can already picture their commercial after the “Debit Card” Monopoly hits the stores:

• Board Game: $15
• Movie Rental: $4
• Zipping through the tedious parts of the board game so that you can put the kids to bed and watch the movie: Priceless

Family board games are all about the process of playing and spending time together, not Tivo-ing through the gameplay to get to the end. Plus, the swapping of cash among family members is a time-honored tradition. As well as a great (achoo!) way of teaching the youngins to wash their hands after handling money. You never know who touched it last.

At the end of the day, if little Mandy doesn’t grow up learning how to manage her own cash in the cocoon-like setting of the typical family Monopoly game (cheating, temper tantrums, and wicked accusations aside), she’ll end up like the other Bratz dolls dancing at the Community Chest.

You know, it’s that charming little joint at the corner of Baltic and Mediterranean Ave.

And evidently the last place in Atlantic City where Monopoly money still reigns.

Of course, that’s just my opinion, I could be wrong.

(With obvious apologies to Dennis)

Also, with apologies to Canada. I had never heard of the Monopoly/Canadian money analogy before this morning. But the line popped into my head and I had to run with it, eh?

How did The Little Ol’ Company From Redmond set the project management world back decades?

They named their product Microsoft Project when it should have been Microsoft Tasks.

This marketing coup has has a profound impact on the business world, because many people now think that a project plan is a file with .mpp extension. The focus on Task Management at the expense of Risk Management has profound implications on the way projects are planned and executed. However, it’s Microsoft’s implicit market dominance, not some evil conspiratorial plot, that has inadvertently led the world astray.

Let me give you a quick, hypothetical example. Let’s say that Wal-Mart grew to the point where they were responsible for 90% of all retail business. In this world, Wal-Mart isn’t a place to shop, it’s the place to shop. When people think “store”, or “shopping”, they equate it with Wal-Mart. In fact, a majority of people have never seen a store that wasn’t a Wal-Mart.

Given this virtual monopoly, Wal-Mart then goes and builds out a very successful private label line of products. Lots of people shampoo with with WM CleanHair, kill insects with WM BugZap, etc.

Unfortunately, when they release their new private label wood stain, they call it WM WoodRefinish. Their market dominance has most people equating the job of furniture refinishing with the use of the WoodRefinish product. Heck, lots of people buy the product, and it turns their wood table into a slightly darker wood table. Success-amundo in their eyes.

Now you have lots of people who think that refinishing a piece of wood furniture is a matter of using a bottle of WM WoodRefinish. They don’t realize at first that stripping, sanding and surface prepping are the important parts of a good refinishing job. Maybe that’s why over 80% of all furniture refinishing jobs seem to take longer and yield poorer quality than expected.

But it’s not the fault of the WM WoodRefinish product. It’s perfectly good wood stain. Plus, the focus groups and marketing folks thought that “WoodRefinish” was a good name for the product.

The catastrophic failure lies in the false perception of the consumer that “applying stain” == “refinishing wood”.

Which brings us back to my central argument: Project management is not about task management, it is about risk management. Using the above analogy, the Gantt chart is the wood stain, but dependency analysis and risk management is the sanding and stripping. They are the time-consuming, labor-intensive precursors to a realistic, good looking Gantt chart.

This is why ProjectPipe is Risk-based, rather than Task-based. It intentionally isn’t centered around building Gantt charts, instead integrating with MS Project for Gantt chart creation.

Rather, ProjectPipe focuses on helping you build links and dependencies among your requirements, issues, tasks, etc. Once you understand how everything fits together, you will have a much better sense for the real tasks that you need to execute and the milestones that demonstrate real progress.

Using a risk-driven, dependency-centric approach, your task list and risk list probably won’t be perfect out-of-the-gate, but they won’t be arbitrary either.

Or, you could just just bang out a list of tasks, guesstimate their duration, wire up their finish and start dates, generate a pretty picture of cascading rectangles, then kick back and call it a day.

It’s up to you.

Our August newsletter is now online.  It is our first issue.   In addition to covering company news and ProjectPipe updates from the last couple of months, we also provide a sneak peek at our new hosted data syndication application, AtomAntenna, which will be in beta shortly.  Check out the newsletter for more info.

If you own a Mac and haven’t tried Quicksilver, you may be doing yourself a profound disservice. Quicksilver is a smart application switcher / desktop workflow tool. Here’s a better explanation.

I’m pretty much using it for doing simple stuff like opening apps and finding files, but it seems that once you achieve Quicksliver guru status, you could rip through reasonably complex multi-step workflows like kids pull off long Tekken combos, in that the heavy lifting can be relegated to muscle memory.
After using it for a few weeks now, I’d have a real tough time getting rid of it.

I just installed the Cube interface, and although it’s arguably just eye candy, that UI seems more Quicksilver-ish (or true to itself) than the other interface options. Here’s a movie of the cube in action, recorded by Tim Gaden at Hawk Wings.