Category Archives: Development Process

perpetual novice – years of experience vs skill

Posted on by
10

Most job listings ask for a certain number of years of experience.  Either in a particular technology or in the industry.  This presumes that the developer learned each year.

I read a quote about whether the candidate has five years of experience or the same year of experience five times.  I don’t remember where this came from, but I think it clearly expressed an important point – that we can’t just measure years.

Upon re-reading the “Software Craftsmanship“, I came across the concept/phrase “perpetual novice.”  This is exactly the idea.  If one never picks up best practices and software development habits, all that time could be spent repeating inefficient ways of doing things.

I’ve seen this upon interviewing as well.  Some people with “X years experience coding Java” look like they can’t code their way out of a paper bag.  I’m thinking this is related to the “perpetual novice” concept.  These people figure out everything from scratch and repeatedly solve problems the industry has surpassed.

What’s dangerous is that such developers don’t consider themselves novices anymore.  Someone new to the field or right out of college realizes they are a novice.  They ask questions and realize there is much to learn.  “Perpetual novices” think they know quite a bit and are less likely to ask a more experienced person.

And worse is when the real novices learn from perpetual novices – the opportunity to learn better ways is lost and the cycle continues!

Have you encountered any perpetual novices in your career?  What do you think makes them stand out?

process maturity and the schedule

Posted on by
Reply

“When we started our project nobody bother about <activity A> because we had very tight delivery schedule.”

Every so often, one sees a statement of this form.  I think it points out that “Activity A” is not an essential part of the process.

It could be because ‘Activity A’ isn’t perceived as having enough value yet.  I think it is a sign of process maturity when this statement doesn’t occur.  For example, I remember a time when we threw tests and the automated build out the window for emergency bug fixes.  Over time, we appreciated the value of these tools and especially the safety net they give us.  I don’t think anyone has thrown out these parts of the process in years!

What really hits me with the quote is that, ‘Activity A’ is perceived as having some value since the author of the quote feels compelled to justify why it wasn’t done.  This is like justifying not brushing ones teeth!

As a process matures you get different levels of value:

– not there yet – don’t even feel the need to make excuses

– immaturity – do the activity sometimes but throw it out the window when things get tough

– maturity – do the activity always

This activity could be anything.  Seeing it get done for emergency production fixes really shows me it has been fully embraced as part of the process.

Subversion, Subversion – what client shall I pick?

Posted on by
12

Since I’m using Eclipse at home for development, it seemed logical to go with Subclipse or Subversive – two of the top three clients listed on the Subversion homepage.  I also tried TortoiseSVN the the other of the top three.  Below are my experiences with the three.  (I haven’t tried branching yet in any of them.)  All have had a release within the past month.  My comments for Subclipse and Subversive are largely from slightly earlier versions though.

Note: See If you tried Subversive before, it’s time to try it again for an update.

Subclipse 1.4.8 (released February 27, 2009)
Pros:

  • It seemed faster to checkout a large project with Subclipse than Subversive.  I can’t prove this as my internet connection varies, but I did notice this a few times.   Another person had the same observation though.
  • Has been around longer.

Cons:

  • No tag awareness!  When I want to commit a file, I need to think about copying the directory to the tags directory.  This is a mental jump as I’m really just thinking I want to create a version – not about how Subversion data is stored.  I also worry that it is too easy to accidentally create the version in the wrong place.  Why invite trouble?  The CVS Eclipse plugin just asked for the name of the tag.
  • Similarly, I have to think about where the tags are stored when comparing to a past version.
  • One common thing I do in Eclipse is awkward – I want to compare the contents of a file to tagged versions.  If I do “compare with revision” I see all the revision numbers and commit comments, but the tag column is blank.  If I do “compare with branch/tag”, I see all the tags, but can only select one to see the compare before starting the whole process again.  Also, you can only compare with branch/tag on a project level – which takes forever – at least over my internet connection.  As a clumsy workaround, we’ve been storing the tag name and revision number on our “release notes” wiki page.  This is an extra manual step that I feel should be unnecessary.  If the tool would just show the tag in the “compare with revision” view like Eclipse does, it would be perfect.
  • The proposal to become an Eclipse project has been withdrawn.

How long I used before switching
A few months.  It wasn’t until the project got into the tagging phase when the tag awareness feature became problematic.

Subversive 0.7.7 (release February 24, 2009)
Pros:

  • Can compare file with “working copy” (last checked out) without connecting to network
  • Since November 2007, has been an incubator project on eclipse.org.  This reset the version number which is why .7 is newer than 1.0/1.1 on polarion.org.

Cons:

  • Installing requires two parts.  The connectors are a separate install for licensing purposes.
  • As in Subclipse, I want to compare the contents of a file to tagged versions.  Using “compare with revision” is worse than in Subclipse.  it doesn’t attempt to show the tags.  More importantly, you must pick which revision you want to compare to, wait a while, view it and then repeat the entire process to pick another revision.  Not nearly as easy to use as the Eclipse version.  (If you think to open the SVN history view first, you get the Eclipse/Subclipse based view where you see them all at once.  It still takes forever though for large projects as it is seeing what else changed in that revision.  Turning off “deep copy” didn’t help matters.)  I didn’t try compare to with branch/tag because I didn’t see the option.

How long I used before switching
I started with Subversive.  After a few months a switched to Subclipse.  Then I switched back to Subversive where I have been since December.

I wasn’t thrilled with either of these so I then tried a standalone client. Tortoise SVN 1.5.9 (released February 27, 2009)
Pros:

  • Checking out one project wasn’t hard.
  • I like the GUI – very intuitive.
  • Committing was simple from Windows.
  • A very cool graphical release view for tags and branching.
  • I really like the log viewer letting you filter by date and or message.  It makes it easy to find out which version you want to compare with.

Cons:

  • It’s not in the IDE.  If not using an IDE, this isn’t a problem of course.
  • Windows only.
  • To tag a project, you need to be aware of the directory structure.  (See my comments about Subclipse for why this scares me.)  There is a graphical explorer to find the tags directory.

How long I used before switching
I just tried it for a few days.  I really wanted an Eclipse based plugin – was just looking to see if this was better.  If I wasn’t using Eclipse, this would be fine.

Conclusion

Both Subclipse and Subversion are usable Eclipse plugins for Subversion.  I prefer Subversive a bit and am going with that at home.  The main reason being tag awareness (I tend to deploy/tag on a weekday evening when I am tired and more likely to mess things up.)  I am keeping TortoiseSVN on my machine for the non-Java code checkins I need to do.  Note that you do have to pick one or the other.    You can’t point to both Tortoise and an Eclipse plugin for the same directory.  As a result, the last week of development on my home computer for Javaranch looked like:

  1. Sunday – Do a bit of Java development in JForum project using Eclipse.  Check out/commit using Subversive.
  2. Monday night – Generate the book promotion materials in PickWinners project by updating the build.properties and running an Ant build script.  (I don’t check these in on Monday since they are so easy to regenerate if something happens to my computer over the course of the week.)
  3. Wednesday night – Production deployment.  I wanted to get in some changes that would make sending private messages to the winners a two click operation.  (It took six clicks for each winner last time it was my turn.)  Tag JForum project using Subversive.
  4. Friday night – Pick winners for book promotion copying winners from web page into one of the files generated on Monday.  Commit using TortoiseSVN.  I like that I didn’t need to open Eclipse on Friday – picking winners is now a web page and the commit happened in Windows.

Many weeks I do less Subversion work at home.  It so happens last week was representative of the breadth of SVN operations I do.  I liked the split between Subversive for JForum (Java development) vs TortoiseSVN for PickWinners (Ant build and text files) and plan to continue that way. I’ll be interested to hear what my co-promotion coordinator uses to commit when it is his turn.  I’m not overly thrilled with either Subversion Eclipse plugin.  I’m thinking of pulling in the JForum project into another directory in Tortoise to use it for file comparison too!

This analysis isn’t so useful for recommending a tool in a corporate scenario where network connections are more reliable and projects are huge.  My suspicion is that Subclipse is better in that environment.  Especially if the majority of tagging operations are done through an automated build.