-4

Good Friends and Good Times

2005-10-22T21:52:00.000-07:00

You know everything is alright in the world when you can call up a few friends you haven't talked to in months, meet up for a beer and have a great time. Beer, Wings, Laughs... fantastic!

My favorite watering hole is called BW3... well.. it used to be called BW3. Now its called Buffalo Wild Wings. I've been going with these guys for more years than I care to say. Though over the last few years its been pretty seldom (maybe twice a year). Tonight was a great night!

Only in Minnesota

2005-10-09T20:43:00.000-07:00

You gotta love Minnesota weather. Yesterday it was over 70 degrees and extremely humid. This was followed by a couple of inches of rain over night. Within 24 hours, we've got 35 degrees (with only 40 for a high) and snow flurries. It was in the 80s over the weekend and is likely to be in the mid-70s early next week. Yesterday I drove to work with the A/C on and drove home with the heat on. The weather in Minnesota is rarely boring. :-)

I saw this quote the other day

2005-08-14T21:12:00.000-07:00

"See if the law takes from some persons what belongs to them, and gives it to other persons to whom it does not belong. See if the law benefits one citizen at the expense of another by doing what the citizen himself cannot do without committing a crime. Then abolish this law without delay..." Federic Bastiat, "The Law", 1850

Good food for thought....

GB-PVR is Awesome!!!

2005-08-14T20:50:00.000-07:00

I'm up and running with GB-PVR and its working very nicely. I installed SageTV and it might have better "TV Guide Searching" capability than GB-PVR, but... I couldn't get my remote to work with it (Hauppauge) yet so I stopped messing around with it. I will try it again, but...

I found a huge "Impress the Wife" feature. It may have won her over to funding a new PC dedicated to being a PVR. I downloaded the Weather plugin (Very nice features, but it seems buggy and when it errors out it takes away all keyboard input from GB-PVR until I restart the GUI). But, the key thing was I downloaded the "My Pictures" plugin. It will do a random slideshow of all the pictures in your picture library (a set of folders you've told GB-PVR about). I combined that with the "Music Library" feature and had a nice slideshow of our pictures with Moby playing through the TV's surround sound audio. It was VERY slick. That plus all the obvious other features might have just won the day for the homebuilt PVR!

1st Phase of PVR DONE!!!!

2005-08-10T19:18:00.000-07:00

I installed my Hauppauge PVR-150 last night. I was all excited. I'd waited all day to do it. Wife was going out for the night so as soon the kids were in bed I cracked open the case! The card itself was as simple as any other PCI card to install. I'm attempting to run it on our "family PC" for now. I want to make sure it works, see if I like it and if my family will like it, then build a PC specifically for it.

So... I connected the cable to the tuner card and put the covers back on and hit the power switch and.... NOTHING. It was dead!

I guess powering it down for the first time in weeks and adding another PCI card to the mix was enough to push the (cheap) power supply over the edge. So... when my wife got home I had to explain to her why she wouldn't be able to use the PC. :-(

Today I swung past CompUSA (that place drives me nuts, but that's another topic for another day) and picked up a new power supply. Once I installed that... I was in business!!!

I installed the software that came with the Hauppauge and then installed GB-PVR. GB-PVR worked VERY well. The remote that came with the Hauppauge worked well with it two (once I fugured out that I had to plug in the IR reciever)

The UI seems a little sluggish... not sure if that's my hardware (Duron 1.2GHz) or GB-PVR's fault. The CPU meter doesn't seem to be pegged at all, but I don't think its always very accurate. Still despite being a little sluggish... it is VERY usable and within minutes I had my TV Guide loaded and was recording shows and pausing "Live TV." About the only thing missing now is audio cables needed for playback on the TV. The sound (for now) comes from the PC's speakers.

I recorded a show for my wife and then she watched it, instantly skipping over commercials (not using the "commercial skip" plugin, but just jumping in 60 sec intervals).

I'll probably try BeyondTV and SageTV at somepoint, but... I'm so impressed with GB-PVR right now I think it'll probably be what I stay with.

First Step Taken toward my own PVR!

2005-08-07T20:36:00.000-07:00

I've taken the first (and most important) step in building my own PVR. CompUSA was running a sale the Hauppauge WinTV-PVR 150 for $60.00. So... I picked one up. Haven't had a chance to install it yet and I'll write more after I do. I've decided to pass on the idea of setting up a MythTV/Linux based box - from what I've read it looks like I'd need to be more comfortable with Linux than I am (or have TONS of time). I'm going to go the Windows route with GB-PVR (maybe try Beyond-TV or Sage too). First step is to see how it all works just on the PC. Then I'll move to running it through the TV. Baby steps...

Thinking of Building my PVR

2005-07-23T19:36:00.000-07:00

I've been thinking about building my own PVR (personal video recorder). I don't have Tivo, but I've seen how it works and I think it would be fantastic to have. I'm reluctant to buy a Tivo for several reasons: A) The monthly fee you have to pay, B) Their control over how you skip commercials C) Their data collection of what you watch.

So combine those three issues with my natural tendency to tinker with all things technical and I've a sincere interest in building my own PVR.

I'm interested in MythTV (a Linux based PVR). Building one of those looks really nice. With the KnoppMyth (Knoppix + MythTV) it looks like its fairly straight forward. But, this IS Linux we're talking about and I'm a little concerned that it might take more time than I've got. I'm not much of a Linux user yet Plus, using Linux there's NO chance in using any of the software that would come bundled with the TV Capture card itself (though... truth be told.. its probably not that good anyway).

The other option would be to buld a Windows based PVR. (not a "Media Center" PC though.. that again looks like too many limitations to me) [Sidebar - one of the problems with commercial media tools is that they seem to not be build with 100% of the user's needs in mind, but instead seem to cater to the interests of the media companies. Its understandable, but its NOT what I want.] BeyondTV (with Beyond Media) looks very interesting, but its kinda pricey.

GB-PVR looks very promising as well. Its priced right (Free) and includes most of the functionality of MythTV (they look very similiar).

My thinking is to give KnoppMyth a try, then GB-PVR and fall back with BeyondTV if I have too.

First things first, I have to convince my wife that we "need" a PVR... :-)

I'll post details about my PVR experience when (er.. maybe if) I start...

I'm back

2005-06-14T19:32:00.000-07:00

Well... after three months off I'm back to blogging with Blogger. I've decided to make this blog not about developement. I'll still be posting technical topics (hey.... I am a geek after all), but I'll also be posting other topics of interest.

You can find my professional/development related blogs at: http://www.jroller.com/page/GearDaddie

FanClub stalled out...

2005-03-23T19:01:00.000-08:00

Well... with all the talk of working on FanClub, "real" work has gotten in the way. Since I've last blogged, I've moved into a new house, incorporated and found a new contract (possibly TWO!). Busy... busy... busy...

Also, from here on I moving my blog to JRoller in the hopes of getting a little more exposure.

Check out my blog at: http://www.jroller.com/page/GearDaddie

FanClub: Technical Architecture - redux

2005-02-11T18:52:00.000-08:00

I've been reading up on Tapestry lately and its piqued my interest. I think I'll use it for the UI layer instead of Velocity + SpringMVC. Tapestry appears to be much more OO/Component oriented than the more "traditional" approach. Furthermore, its markup doesn't interfere with your HTML. This allows for WYSIWYG HTML editors to show realistic previews. There's a little higher learning curve with Tapestry, but my hope is that it will pay off in the long run. We'll see how it goes.

FanClub: Initial User Roles

2005-02-11T18:42:00.000-08:00

Following the process outlined in Mike Cohn's book "User Stories Applied" I've come up with an initial list of "user roles" from which I'll derive "user stories."

Avid Fan (think Trekkie)
Casual Fan
Gift Buyer
Member Manager
Order Manager
System Administrator
Business Owner

Of these, the most important roles to the success of the system is probably the Avid Fan and Order Manager. The other roles are all important and will be included, but when it comes to priorities, those two roles take center stage. The definitions of these roles should be fairly obvious so I won't detail them here. More specific details will come with descriptions of the User Stories.

Don't leave commented out code in your source!

2005-02-08T18:44:00.000-08:00

Quite often while digging through code I'll run across blocks of code commented out. Usually I have no idea why it was commented out, nor why it was left there. Regardless, whenever I see it, I immediately remove it. Source code with blocks of unused code is a huge sign of code rot.

I think many developers get in the habit of commenting out code because they think they might need it later. Or worse... aren't entirely sure of their changes and don't want to lose the original. If their solution seems to work, they release the code and expect that they'll probably come back to it if people report problems.

Don't do it! As soon as you've tested your code (you are testing, aren't you?) remove the commented out code. Your SCM is your back up. With a good SCM (Eclipse + CVS for example) and an IDE with "Local History" (did I mention Eclipse? ) you can always recover old code.

Playing to not lose.

2005-02-05T19:55:00.000-08:00

I've worked for a few organizations now where they're are putting a lot of emphasis on process and review. It takes forever to get anything done in these organizations. Money and time are wasted as people push projects around in the process. I think the primary reason these companies are doing all of this is to protect them from runaway projects that end in disaster. I can certainly understand the motivation and have witnessed first hand some of these types of projects.

What bothers me is that these processes are setup as though they are "playing not to lose" instead of "playing to win." They slow teams down and take all of the initiative away. Leaving most with a "who cares" attitude. And why shouldn't they? The company has put their faith in the process to see them through to success instead of in the people. The company then can't figure out why they can't get the responsiveness from their IT department, they wonder why its so expensive. In the end they'll decide to outsource it to some other company that can do it quicker and cheaper! (who, of course don't have nearly as involved process or who do have the process, but offshore it for cheap labor).

In all the projects I've ever been on, it has always been the people who made the project successful. Every time the process (created by another group) has hindered us delivering value to the customer.

Don't get me wrong... every project/team needs some process. There should be corporate standards. But, they all need to be justifiable in terms of business value and should error on the side of being too little, rather than too much. All that should be done is: hold the team accountable for the success (and the management above them).

Hire the best you can find and trust them.

Apple's Arrogance

2005-02-05T19:38:00.000-08:00

I was thinking the other day about how I'd love a 15" PowerBook, but after I read this thread on another blog I'm having second thoughts about it until the G5 PowerBooks come out (which may never happen). It seems like everyone is in love with their PowerBook, but yet they are slow. I just can't see shelling out that kind of money for a laptop that's not up to the task for me. (To be fair... I need basically a portable desktop for development - my requirements are quite high ).

If onlyApple would hook up with AMD! Could you imagine having a 3700+ Mobile Athlon PowerBook? Apple has had a X86 version of OS-X floating around their office for years now. Heck, with that kind of power they could emulate the G4 and not lose any performance over their existing hardware. And... a dual 3700 Athlon system would probably be faster than the current dual G5s (and cheaper to produce too!).

The market is ripe for the "switch" campaign more now than ever. And yet... I think Apple's arrogance will never allow them to move to a X86 based platform. This leaves a large number of people (me included) sitting on the fence, wanting an Apple but not being able to justify the price/performance.

That said... I'm still really interesting in a Mini-Mac, but... that'll never go to work with me. :(

Microsoft Project.... run away!

2005-01-31T20:34:00.000-08:00

Every project I've been on lately has had a project manager who lived by Microsoft Project. For the first 2-3 weeks of the project he/she would spend nearly 100% of their time working on getting the project "setup." I'd give estimates for my tasks and so would everyone else. Then he/she'd run off printouts every week that would show the schedule.

But... after all this effort:

it was always out of date
it was never accurate, but overly precise (saying that my tasks would be done in 5.15 days) giving a false sense of confidence
is was brittle and eventually everyone avoided correcting their estimates (it didn't matter anyway.. see 1, 2)
it was always abandoned later in the project for something like a spreadsheet

Why do project managers still do it? I think it mostly has to do with a "CYA" attitude. Its as if they can go back to the stakeholders and say.. "See... I planned to have it done on time. If they'd have just stuck to the plan we'd be OK - I've done my job..."

I don't see any value in having things put in MS Project (or any other tool of that type). It has never helped me get a sense of how the project is moving along, nor have I felt like its reflected how things were going for my tasks on the project. Things just seem to get lost in the noise.

I believe the Agile project management practices have a lot going for it. Short iterations, tracking velocity, burn down charts, etc are all nice and simple. And, more importantly give a much better eye into how the project is progressing.

Keep it simple and post the information you do track on big posters or whiteboards so everyone can see it and change it.

Better Pet Shop - Tapestry, Spring, Hibernate

2005-01-30T10:12:00.000-08:00

I found this link the other day and it may be useful as I work on my FanClub project. Its an implementation of "Pet Shop" using Spring, Hibernate and Tapestry. Though.... it looks like it hasn't been modified in a long time. Could be a dead project...

Java & .NET

2005-01-29T20:12:00.000-08:00

I found this link today and found his thoughts very interesting. I'd have to say I share his thoughts on the productivity of Java. I think it CAN be productive, but the way many organizations are using it.. it is NOT very productive. I've used .NET enough to know that if you want to build a typical web app ASP.NET is a pretty productive environment. It is also pretty much the only way to develop web apps in .NET. I've wondered, as he has, if all the layers, patterns, frameworks, etc that are found in typical J2EE applications aren't adding way more complexity than they're worth. After reading his blog... I'm inclined to look more at Tapestry and Hibernate.

FanClub: Technical Architecture

2005-01-28T18:50:00.000-08:00

FanClub is a typical dynamic, data-driven web site. A lot like (too much probably) the typical Sun PetStore example app. I certainly could implement it using a high-end J2EE app server, Struts, EJBs and JSPs. But, instead, my intent is to use a "lighter-weight" solution and use a couple of technologies that aren't as common. I believe, in the end, that this will provide a better solution appropriate to this type of application.

The app will be divided into the typical logical layers of UI -> Business Logic -> Persistence all running within the same app server (a servlet container). Following the basic design that Matt Raible uses with his Equinox starter app, each layer will talk with the other via interfaces.

The application will rely heavily on the Spring Framework to tie everything together with its Dependency Injection. It hels enforce good design, encourages "design to interface" and simplifies testing.

I also intend to use the SpringMVC framework for the "controller" part of the UI. SpringMVC is more powerful and flexible that Struts and since its there... I'll use it. For the "view" part I'm going a little off the beaten path an using Velocity. Why Velocity? I've never been much of a JSP fan. The syntax makes viewing it in an HTML editor messy. I also don't like the JSP -> Compile to Servlet process. Its never felt right to me. Velocity is fast, has a simple syntax and doesn't interfere with HTML.

For the persistence layer I intend on using iBatis. iBatis is a very simple object-relational maping tool. Since I expect my database interactions to be fairly simple and iBatis is easy to get use with I'll start with it. I may move to Hibernate later on. Following a DAO pattern for persistence I should be able to swap iBatis for Hibernate fairly easily.

Finally for security, I intend on using the Acegi security framework. It integrates nicely with Spring and has no container specific dependencies.

The application server for this project will be Jetty. Jetty is fast and light-weight. Tomcat would work as well, but my experience with Jetty has been that it is faster to develop with and peforms just as well under the loads expected of this app.

Also, as long as I'm at it, I plan to use Java 5.0 (JDK 1.5).

For tools I intend on using:

Eclipse 3.1M4 (with JDK 1.5 support) - IDE
Jetty Launcher for Eclipse (an excellent plugin for managing Jetty) - IDE
Velocity Web Edit (Eclipse plugin providing HTML & Velocity syntax support as well as help with "soft references" in the pages) - IDE
CVS - version control
Ant - for builds
JUnit - testing
FIT (or Fitnesse) - acceptance testing
Hypersonic DB (HSQLDB) - RDBMS

Two cool apps...

2005-01-28T18:40:00.000-08:00

First, feeding my Mac envy I found an Expose' like utility for Windows. Its called TopDesk. Its pretty slick. I've only tried the "free" version so far, but for $10.00 it seems like it might be worth it. It takes all your open windows at tiles them for you (or zooms out, if you will) , then it lets you select one then "zooms" back in to it. Check it out.

Also, my other handy app is called VirtuaWin. It creates "virtual desktops" for Windows. Its much better than the Power Toy you can get from Microsoft, though not as fancy as others. Its open source and free! I have a 2 X 2 desktop setup both at home and at work. I have it configured to navigate to the next desktop when I hold the Ctrl key and move to the edge of the screen. By default it does it after you've paused at the edge (no Ctrl keypress required) and moves the mouse on you. That setup didn't work for me.

One note, these two apps seem to conflict with each other, so you'll have to pick one or the other. :-S

FanClub: Project Vision Statement

2005-01-28T07:38:00.000-08:00

To start the project, I've come up with a Project Vision Statement. Nearly all methodologies recommend having a project vision statement. It is intended to help guide the project and help in making decisions as to what should and should not be included in the project.

Here' is the project statement for my project: FanClub

FanClub's goal is to serve as an online store where members can buy branded FanClub merchandise, find out the lastest news about the FanClub, and communicate with other members.

Next step, identifying the high-level architecture and technologies...

New Project

2005-01-27T20:29:00.000-08:00

I've been kicking around the idea of doing a pet project for some time now. I've been wanting to do a project using agile methods and light-weight J2EE architecture. I can't seem to find one yet (if you've one let me know! ). So, with that in mind I've decide to take an old database application I wrote 10 years ago (man... am I old) and convert it to a web application. The original application was for keeping a list of club member, sending them newsletters, and processing orders for branded items sold to the club members.

My plan is to create an "online store" (yes... another Pet Store type application) with news and possibly discussion forums as well.

I intend on following a "Test Driven" development and using "User Stories" as a basis for directing the project. Though its just me doing all of it, I hope to make it a good learning experience. I thought I would share my experience here on this blog. Stay tuned...

(of course... I'm moving in the near future so I expect that I won't have a lot of time over the next couple of months)

Trillian

2005-01-27T20:09:00.000-08:00

I've been a big user of "chat" programs for some time. I started with ICQ and then moved on to MSN Messenger. Over time, of course, I found myself ending up with friends on AIM and Yahoo as well. Running multiple clients was a pain and just didn't feel right to me (too much duplication I guess).

I'd tried Trillian. It has all the features I want (mostly that it can connect to all the major Chat servers). But, each time I tried it I found myself feeling like it was a "least common denominator" and the UI was too gimmicky. I eventually started using GAIM and its "OK", but its UI is the opposite of Trillian... its VERY plain and really feels like the least common denominator.

Things change.

I was having dinner with some fellow collegues of mine and one of them mentioned Trillian and its "new" release.

I've been running it for a little while now and I have to say... its great! I highly recommend it. I'm thinking about forking over the $25 for the "Pro" version. The UI is much improved and the very polished (if still a little gimmicky). Check it out! Let me know what you think.

The mini Mac

2005-01-14T19:22:00.000-08:00

As you may or may not know Apple computer has released a new Mac. From my perspective its a very smart move. I've always had an appreciation for Macs. I had an original Mac with 128K back in 1984 (I still have it someplace...) and I remember it fondly.

As a Java Geek, I've thought about switching over to a Mac. It has all the Unix "goodness" plus a slick, attractive UI (and I'm a GUI nut). I've really wanted to try OS-X. I see a lot of Java developers (or at least Java presenters) using Macs and it seems to work well for them.

But... it would be a bit of an experiment to adopt a Mac and I'm not sure I'd want to spent that kind of money on a PC and find out its not right for me. But... at $500 (or a little more for a RAM upgrade) it seems palatable to me. If I end up not liking it for some reason... it'll still work as a nice Internet, Photos and Music PC. If I end up really liking it... who knows, maybe I'll end up with a G5 iMac or PowerMac. :-D

So... from my perspective I think its a smart play for Apple. It allows a lot of us "sensible" geeks to try the Mac without a huge investment. The more geeks Apple can get, the more successful they'll be as geeks tend to drive the market.

What Subversion needs...

2005-01-08T19:58:00.000-08:00

... (well for me at least) is the type of integration built into Eclipse that CVS has. I believe if it had that level of integration it would rapidly replace CVS in a lot of shops. The CVS support in Eclipse is incredibly full featured and allows for near seamless usage. There is a project (subclipse) that appears to be working toward that same level, but progress has been really slow.

Poor progress on "the list"

2005-01-06T19:40:00.000-08:00

After talking about Subversion, CruiseControl, Maven, Fitnesse and Equinox... I've only had time to look into Equinox. If you're looking for an easy way into learning Spring, I would definately check it out. It took me 10 minutes to get a working version of Equinox setup and I just started adding what I wanted to it and removing what I didn't need in small steps. A great learning tool! Hats off to Matt Raible and co.

I found Spring to be fairly easy to use and I experimented with the iBatis persistence framework and Velocity as a view technology instead of JSPs.

iBatis seems very useful if you have a "database-centric" application as it allows you to use the power of SQL easily and its got a very small learing curve.

Velocity works nice as a view technology because its easy to learn, doesn't corrupt your HTML like JSPs and needs no compilation step. Using Jetty as my container made for really fast development.

My next step, I think, will be to figure out how to make use of Fitnesse.

Version Control and Branching

2005-01-06T18:20:00.000-08:00

Today I found that our CA (configuration analyst) put a print out of a link to somebody's opinion of how long it takes to merge changes once you've branched. In it, it states that to calculate how long it takes to merge changes based on the length of time the branch existed. To calculate, just take the of units and add a "unit of measure." For example, 1 day of branch work takes 1 hour to merge, 1 week -> 1 day, etc.

During our last project the CA was livid when it was decided to branch (he's the guy who has to manage the SCM system) and yet it gave us the flexibility we needed.

We branched because we had two different "feature sets" that might need to be released independently. The code sets were 95% independent of each other. In the end he claimed it took 7 days to merge the code after living with the branches for 8 weeks.

I was very surprised it took this long. I've (for the most part) had little trouble merging code on previous projects (using CVS mostly). In our case, I believe it had to do with the tools (Clear Case), their usage of it and the complexity of their architecture made dealing with the changes hard. Furthermore, the merging of the code was the responsibility of the CA instead of the developers that wrote the code. That definately has to be an anti-pattern...

What are your thoughts? Has branching and merging been huge problems for you? What SCM tools have you used and found helpful?

If you're interested in Agile development...

2005-01-05T21:13:00.000-08:00

Check out Alistair Cockburn's new book Crystal Clear. I read it over the last couple of weeks and found it very insightful. (especially chapters 2,3 and 4). Now I just need to find a client that wants to adopt agile development...

Uninspired by Linux

2005-01-05T21:01:00.000-08:00

I've got Fedora 3 installed and its quite nice and yet... I find myself gravitating back to XP. I'm not sure why. I want to like Linux more. I'd like to replace it as my "main OS." But it just hasn't grabbed me yet. I'm a stickler for user interfaces and Linux is definately behind XP, though not by a lot. I love the tabbed terminal windows and the power of the BASH shell, though I'm not a big "shell" guy. I like Evolution, but it feels like Outlook 98 in many ways. I like Rhymbox for playing music, but not nearly as much as MusicMatch or Media Player 10.

Certainly as a server in my home setup it would be great. Put JBoss, Tomcat, CVS, Subversion, etc on it would work nice. But, as a workstation, I still enjoy the "experience" of using Windows (no pun intended) more at this point. Perhaps if I had two boxes instead of having to dual-boot I'd use Linux more. I don't know...

I have a feeling Mac OS-X 10.3 is probably nirvana for me, but I'm too cheap for that. My next PC needs to be a laptop and Powerbooks are expensive and slow. I wish I could try OS-X for awhile with out such a big investment. (perhaps the rumored xMac?) Oh well....

Am I missing something with Linux?

Taking Architecture and J2EE too far...

2005-01-05T20:00:00.000-08:00

A read-only web application that reads from a single database that does little processing (90% of the logic is in selecting the data and massaging it for the UI) uses over 6 separate EJB heavy "enterprise applications" consisting of one application with several "business service providers"? What's going on here?

What's really interesting to me is that none of the techincal decisons were challenged. There was no evaluating the chosen architecture versus other alternatives. And even though it costs us daily in wasted productivity it appears that most everyone just accepts it as "the way it has to be". I checked one day and even a relatively simple request from the user traveled through 11 EJBs (most using Remote Interfaces) across 3 separate EARs. Do you think there could be a performance problem?

Our industry as a whole needs to get much better at the appropriate use of technology. In this case there is no reason for physical separation of layers, nor the usage of EJBs. Logical layers for this application are appropriate and putting reusable code into JAR libraries makes sense. But for a read-only, non transactional web app... c'mon...

No wonder CTOs are outsourcing. Even if they got the same productivity constraining architecture, they won't be spending as much per hour on it...

Becoming Subverted....

2004-11-18T19:59:00.000-08:00

Yesterday I downloaded Subversion and installed it. I installed a plugin for Eclipse and the Tortoisesvn Windows Explorer GUI. So far I'm impressed with SVN and both GUIs. It was really easy to setup the basic, lightweight server and get my client talking to it. It looks like I won't be missing much of anything over CVS and gaining quite a bit in return. I blog more about how it goes as these are just my initial impressions. Now I've gotta get CruiseControl working with it....

Next up on the list...

2004-11-15T19:08:00.000-08:00

If I ever get the time (winter is coming to this part of the world so it is possible), I'm planning on trying out: Subversion, Equinox, Fitnesse, Maven, and CruiseControl. I think all of these tools could be VERY useful in upcoming projects. I plan on leveraging any support they have for Eclipse. Did I mention how wonderful I think Eclipse is?

What is up with Ruby?

2004-11-15T18:53:00.000-08:00

There seems to be a full court press by all the "thought leaders" (Martin Fowler, Dave Thomas, etc) on Ruby lately. I know some of it has to do with Rails and from what I've seen, it DOES look impressive (Rails that is). I've tried Ruby a few years ago and was impressed with the language, but at the time the tools/libraries surrounding it were a little thin. By the look of Rails things have improved. I suppose I'll have to buy the 2nd "Pick Axe" book. I used to think that strong-typing was really important, but as I think about it, J2EE development is filled with untyped lists (though JDK 5 will change some of this) and tons of soft-references in JSP pages, config files, etc. Perhaps, with a test-driven development its not a big deal?

Fedora Core 3

2004-11-09T17:45:00.000-08:00

I've downloaded the latest Fedora release and I'm looking forward to installing it. I'll post my impressions of it here. I've used Fedora Core 2 some and I have yet to get hooked on Linux. I'm not sure why. I am a bit of an OS junkie. I should be crazy about Linux. Maybe I'm just getting old...

Don't forget the IDE!!

2004-11-07T20:01:00.000-08:00

Most Java developers I talk to always talk about the importance of being able to build the application using ANT. "I should be able to walk into a project, check out the code, type ANT and have it work." Or for the enlightened folk, replace ANT with Maven. ;-)

While I whole-heartedly agree with this, don't forget the IDE. Ant gives you a LOT of flexiblility with how you structure your source files. This is great, but it also can render a lot of the power of your IDE useless. The most common example is the layout of web applications. I've worked on several projects where the "build master" has gotten creative with the locations of the web files. This has rendered Struts Editors, JSP Editors, hot deployment features, etc. useless. It also has made Cactus, StrutsTestCase and other supporting tools VERY cumbersome to use (or useless).

In the end, I think projects/teams should reign in some of the Ant creativity and stick to standard "exploded" WAR, EJB, EAR formats trying to leverage all that they can from their IDE. Otherwise... its productivity wasted. Or... in other words... if you're going to get creative, make sure the benefit justifies the cost.

A poor SCM system costs...

2004-11-07T19:30:00.000-08:00

I've worked with several SCM tools over the years. I used to tollerate PVCS, then I moved to StarTeam. I really liked StarTeam. Then I started working with CVS. At first I hated it. I was a GUI guy and it seemed so "old fashioned" to be using the command-line. WinCVS was "OK", but I still didn't get it. I didn't have the right mental picture of CVS's "optimisitc locking" and "sandbox" implementation.

But... once I did.... Wow! It was SO nice to use. I was free to make changes where ever I wanted to without anything in my way. With J2EE development and its unavoidable 10,000 files for even a small project (ok... I'm exagerating a little...) it really help. Finally I discovered Eclipse and its CVS support. It is absolutely the best way to work... hands down. It is a highly productive environment that never gets in your way.

Time change, clients change. I'm now with a large insurance company that has bought into the "Rational" suite of software (that name sounds so ironic these days) using ClearCase and the productivity of my team suffers heavily for it.

In particular for one, they insist on "dynamic views" created for each individual developer. This is basically a virtual drive on your PC that is under total control by ClearCase. You can't touch any file on this drive without getting ClearCase's permission. Furthermore, you can't check out a file unless you have a "reason" too. It is a pesimistic locking scheme, but with each developer having their own "dynamic view" that needs to be delivered to the "main view", why do they need to lock it? Makes no sense.

I'm getting off topic a little, but my point is: it is costing my client thousands of dollars a day as developers fight this tool. (Having somebody else give them permission to work on code and get somebody to approve the code so others can see it.) On a really large team on very sensitive software, I could understand it. But with only 5 developers? Its a waste. When somebody raised the issue the response was that "we can get a report on who changed what and why that code was changed." Sure... that sounds great. I'm sure its nice. But... have they ever evaluated the benefit of that report(s) with the cost? It seems highly unlikely. Using CVS would: a) be free, b) integrate MUCH better with their tool of choice - Eclipse c) rarely hinder developers. If they really need that report from a team of 5 developers - just ask them!

Return on Software

2004-11-07T19:12:00.000-08:00

There's a new book by Steve Tockey called "Return on Software." I spotted it at the local Border's the other day. It looks very interesting. It got me thinking about how little attention is really paid to getting the proper return on the money spent on software. I've worked on dozens of software projects over the years and I can only think of a couple that really calculated the expected return on the money invested in the software project. That's pretty sad. Furthermore, as a software developer, I'm embarrased to admit that I haven't paid more attention to that aspect.

Steve's book looks a little dry, and, as he admits, only the examples are Software oriented. The rest of the material is pretty generic. At any rate, I'll be buying this book or "Making the Software Business Case" soon. Its a skill I need to improve on!

And so it starts...

2004-10-31T20:34:00.000-08:00

I've finally become a blogger... now if only I could think about something to write...