Can you spot the difference:
Scala is a serious option for organizations developing on a JVM platform.
The friction and complexity that comes with using Scala instead of Java isn't offset by enough productivity benefit or reduction of maintenance burden for it to make sense as our default language.
In a Model View Controller application, you only ever want to see SQL code in the model.
The domain model should be independent of persistence implementation details.
Using C means that people [...] don't screw things up with any idiotic "object model" crap.
Lately I seem more susceptible to realize how fast the world is changing. Probably I am getting old...
At a recent lecture, Martin Herdina (CEO, Wikitude) showed how the mobile phone changes our lives (and he displayed slides you don't usually see in lectures…)
Once again I realized that the iPhone was only launched in 2007. I think it is amazing how our use of technology can change in only 4 years ...
Another striking example underlying the perception of technical change in relation to other areas of our lives:
What crosses your mind when you see the following picture?
Old-school, clumsy, clunky, retro …
...at least that are impressions that pop up in my head…
Now, what crosses your mind when you see the following picture:
Cool, highlight, milestone in movie history, top-10-movies of all times, re-defining science fiction movies …
...again that are my personal associations.
But you remember that the Nokia 7110 phone in the above picture was actually designed after being inspired by the Matrix movie:
I think it is remarkable how time is not having the same effect on movies as on technology. The Matrix definitely aged better than the Nokia phone.
In context: Great video about living in exponential times.
There is so much wrong in this setup...
According to a claim in the Non Relational Data Stores Panel, the query parser in MySQL alone spans 100'000 lines of code in C and the whole Cassandra Database is 30'000 lines of Java.
I don't know what to deduce from that, but it is certainly interesting.
Some directions of thought:
Once you've admitted to yourself that you're a bad programmer, you can stop all the silly posturing and pretending that you're great, and you can look around and find the best possible tools to help you look smarter than you are.
People pile layers on top of layers, abstractions on top of abstractions, complications on top of complications, crap on top of patches, and patches on top of crap until everything collapses onto itself and the singularity appears.
In requirements analysis we learn to pop the why stack and to challenge business requirements.
The goal of these techniques is to find the real business value at the heart of a given requirement. This real business value is often hidden by an already envisioned solution. Once we know this underlying business value, we might come up with a better/easier solution. Most often the reason why we can come up with an alternative solution is that we can look at the problem in another context and from different perspectives than the original requester of the feature.
This matches with the concept of 2nd order solutions in systems theory: 1st order solutions work within the existing system while 2nd order solutions modify the system itself.
The classical example from requirements engineering is the requirement that a software system must be able to print each UI-screen. The underlying reason was, that employees manually copied data from the given system into another system by printing them out and then typing them into a terminal of the other system. Once this underlying reason was discovered, an alternative solution could be realized: integrating both systems and digitally export/import data. This solution was cheaper to realize for the implementator and provides more business value for the enduser.
A more trivial example to illustrate this "out of the box thinking" is the following task:
Connect all the 9 dots in a single move, using 4 straight connected lines:
Can you do it with three lines?
Can you do it with one line?
Recently I started studying for MAS-MTEC at ETH Zürich.
A characteristic of this post graduate studies is that most lectures are shared with graduate students.
I am impressed at how much studying today has changed from when I was studying a decade ago. This impressively shows how fast technology is changing the way we live, which I seem to forget in my every-day life.
The internet is ubiquitous when studying today: The studies are officially organized over the internet (lecture selection, examination registration, learning material distribution, task assignments, group formation for exercises ...).
On the first week I got a invitation to a DropBox share from somebody I actually do not know (probably a higher semester student) which contains over 1GB of "semi-official" studying material (solutions for exercises, example examinations, additional material ...)
One lecture even has a twitter hashtag.
Consequently practically every student is bringing a laptop or tablet to the lectures. I never have seen such a high local density of MacBooks as in the Marketing lecture :-)
Comparing that to when I started studying 12 years ago: I did not even own a cell-phone yet ... I also clearly remember when a lecture assistant suggested to use "Google", and I had never heard that strange term ...
In this one day workshop developers can learn how to apply efficient usability-design in their enterprise project.
Check it out: Details and registration.
Some time ago I claimed that the JPA 2.0 metamodel API has the potential to revolutionize Java development.
I still think that the concept is very interesting by showing an approach to strongly typed meta programming in Java. However I think it does not have any relevance in real world projects. One reason is that strongly typed JPA criteria queries are very verbose and bring their own accidental complexity compared with JPQL. The other reason is the actual usage of an annotation processor in any build environment is still too complicated.
In the following I show how to configure the JPA metamodel annotation processor of EclipseLink for different environments. A working example for this is exercise 6.3 in our jpaworkshop.
Maven
EclipseLink documentation is (once again) lacking, ignoring the reality that Maven is currently the most prominent build environment in the enterprise.
Fortunately this is well documented here and here.
You need to configure a maven processor plugin in the pom that triggers the annotation processor in the generate-sources phase:
There are different maven processor plugins. I am using maven-annotation-plugin, an alternative is the Apt Maven Plugin.
Now Maven was easy, lets tackle the IDEs...
NetBeans
NetBeans excels in this task. When opening the Maven pom, it automatically recognizes the above configuration with the maven-processor-plugin and configures itself to use the EclipseLink annotation processor: 
IntelliJ IDEA
Configuration of an annotation processor is nicely documented in this post by JetBrains. 
The good things:
The bad thing:
Eclipse IDE
One might be tempted to think that configuration should be especially easy since Eclipse IDE and EclipseLink imply some kind of close relation...
The EclipseLink documentation explains how to configure the annotation processor ... except it does not work:
In combination with Eclipse the EclipseLink annotation processor does not detect the default META-INF/persistence.xml. You have to configure it manually. This is not documented and not trivial. The problem is reported as bug, but the bug was closed without fixing the problem! I wonder how many people gave up on using the metadata API just because of that shortcoming...
Here is how to configure the EclipseLink annotation processor in Eclipse:
As described in the documentation you have to include three jars on the factory path of the annotation processing configuration:
This approach is bad from the beginning, since those jars might change their name when you update them and they might not versioned together with your sources (this is the case when using Maven) . The approach of IDEA of locating annotation processors in the classpath by their class name is much better in this regard.
But as mentioned above, this does not yet work. You get the following error written in the eclipse error log:
The persistence xml file [META-INF/persistence.xml] was not found. NO GENERATION will occur!! Please ensure a persistence xml file is available either from the CLASS_OUTPUT directory [META-INF/persistence.xml]
The solution is to pass the persistence.xml explicitly to the annotation processor. This is achieved by configuring an annotation processor option. The key is eclipselink.persistencexml. The value is the path to your persistence.xml relative from your CLASS_OUTPUT directory. In case of using Maven, your CLASS_OUTPUT directory is target/classes, so you have to prepend ../../ to your path to arrive in the project root directory ... not trivial indeed ... (note that the path separator might vary on Windows)
Finally generation of JPA metadata API classes is also working in Eclipse. With the great background compilation of Eclipse it is nicely working on the fly when editing files.
Abstracting SQL often isn't a good idea from my experience.
Hibernate should be to programmers what cake mixes are to bakers: beneath their dignity. [...] As professional programmers, we should be more sceptical of generic frameworks like hibernate.
There are easier ways to do it, rather than hitting your domain model over the head with NHibernate.
Documentation for EclipseLink is quite lacking.
While for Hibernate it is quite easy to find out how to log SQL statements, I had some trouble to find out how to accomplish this in EclipseLink.
The solution are the following properties in persistence.xml:
<property name="eclipselink.logging.level.sql" value="FINE"/>
<property name="eclipselink.logging.parameters" value="true"/>
This solution I finally found here.
An alternative is also to use log4jdbc or log4jdbc-remix (the latter is available in the sonatype maven repository). An example is available in exercise 10 of my jpaworkshop.
For my recent MSDN TechTalk I did some research how to migrate a source control repository from SVN to TFS. While I am not advocating this migration, here are the options:
Currently you have a choice of three tools:
svn2tfs is a simple tool written in VB.NET. You need to have SVN installed in order for this tool to work. It basically replays every revision from SVN as changeset into TFS.
The TFS Integration Platform is the Swiss army knife for many possibilities to get data in and out of TFS. The TFS integration platform is available as supported Microsoft product on Visual Studio Code Gallery and as bleeding-edge release on Codeplex. The SVN adapter is currently only part of the codeplex release.
Timely Migration is a commercial tool. There are different modules for different source version control systems. To migrate from SVN to TFS the SVNToTFS module is needed. It costs $1995. Trial versions can be requested, but they do not migrate the content of the files.
| svn2tfs | Integration Platform | Timely Migration |
| open source | open source, from Microsoft | commercial, $1995 |
| simple and easy to use | many features, complex, complicated | powerful features, easy to use |
| enough documentation | sparse documentation | good documentaion |
| SVN tags can be migrated as branches | SVN tags are migrated as branches | SVN tags are migrated to TFS labels |
| user-mapping is enforced | user-mapping is possible but not enforced | user-mapping is enforced and comfortable |
| migration is aborted in case of an error | some errors can be manually resolved and migration can be then be resumed | migration gets resumed when restarting after error resolution |
With each tool I was able to successfully migrate an example project from a local SVN repository.
svn2tfs: I was not able to migrate SVN repositories from Google Code. The tool systematically failed, because the initial repository revision for the project layout (dirs for trunk, tags and labels) has not author. The tool can not deal with that. On the positive side was, that the source code of the tool is very simple so that I was quickly able to create a patch, that fixed this problem. On the negative side was, that there was no reaction on the patch I submitted. The project seem pretty dead.
TFS Integration Platform: It seems not possible to do an aynonymous login to SVN. If you do not provide valid user/password, the migration fails with an ugly “NullReferenceException”. The documentation for the SVN adapter is sparse. For instance there was no documentation at all how to map svn users to TFS users. fortunately (after some nudging from my contact at Microsoft) my question in the forum was answered. Some days later even a blog post about the topic was published. Another shortcoming is, that the mapping from svn users to TFS users is not enforced. If no mapping is specified, each svn user is silently mapped to the user that is running the migration tool. There is no support in checking if all users from svn are mapped to TFS users.
Timely Migration: The only negative thing is the price since you probably need the tool only once.
With none of the tools I was able to migrate the Nerd Dinner repository from Codeplex. Each tool either reported an error or froze indefinitely.
As we are entering the era of language wars on the JVM there is another war looming on the horizon...
This new war is coming out of the realm of distributed version control systems:
Git seems to have a nearly undisputed imperium in this realm and recently has squished its known competitors:
Despite the seemingly striking dominance of Git, there are still brave contenders sprouting out of the dvcs land:
All of them featuring some unique features (although features I am not always sure I would like to see coupled to a source control system ...) and most of them somehow promising more simplicity compared to Git.
I wonder if they have any chance against the massive momentum of Git and its ecosystem ...
There is a common claim that we should learn more from classical engineering disciplines like civil engineering. According to that claim the IT industry would be a better place if we would adopt best practices from the latter.
The amounts of data that Twitter handles is astonishing:
... this is according to the presentation NoSQL at Twitter from October 2010, so today the pile of floppy disks would probably already be 50 miles high...
It is amazing to realize how old the notion of Test Driven Development (TDD) really is. Here are two quotes, that are older than I am, describing the ideas behind TDD :
Report of the NATO Software Engineering Conference 1968:
A software system can best be designed if the testing is interlaced with the designing instead of being used after the design. Through successive repetitions of this process of interlaced testing and design the model ultimately becomes the software system itself. I think that it is the key of the approach that has been suggested, that there is no such question as testing things after the fact with simulation models, but that in effect the testing and the replacement of simulations with modules that are deeper and more detailed goes on with the simulation model controlling, as it were, the place and order in which these things are done.
The humble Programmer, Edsger W. Dijkstra 1972:
But one should not first make the program and then prove its correctness, because then the requirement of providing the proof would only increase the poor programmer's burden. On the contrary: the programmer should let correctness proof and program grow hand in hand. [...] If one first asks oneself what the structure of a convincing proof would be and, having found this, then constructs a program satisfying this proof's requirements, then these correctness concerns turn out to be a very effective heuristic guidance. By definition this approach is only applicable when we restrict ourselves to intellectually manageable programs, but it provides us with effective means for finding a satisfactory one among these.
Today we mostly associate TDD with automated unit-tests and the family of xUnit Frameworks. We associate TDD with writing code.
Yet TDD today is far from being mainstream in my experience. And even teams that decided to adopt TDD are constantly struggling with implementing it in reality.
The participants of the NATO Software Engineering Conference from 1968 probably had quite a different notion of implementing TDD than we have today. Their implementation of TDD was probably far away from the automated unit-tests we strive for today. Yet they had the same mindset that we pursue today.
When Dijkstra was already preaching TDD, the xUnit-Frameworks were not to be born for another 26 years.
In the last few years I have come to the conclusion, that the most important aspect of TDD is the mindset that comes along with TDD.
This mindset is much more valuable than any tool, technology or methodology that is commonly associated with TDD today.
With the right mindset we can practice TDD even in environments where we can't write unit-tests (Lo and behold, I am even thinking of SharePoint Development). TDD does not have to be realized by writing code. Nor is writing unit-test always the best way to implement TDD.
With the right mindset we can even practice TDD with manual testing!
The actual implementation of the test is rather a detail of TDD. Much more important is the mindset of practicing baby steps, the mindset of gaining insights and evolving through rapid feedback, the mindset of leveraging trial & error as a methodology, where errors are not failures but valuable insights that guide the evolution of the project.
This mindset is the reason why children in kindergarden are scoring better than MBA graduates in the Marshmallow Challenge:
And it is our God Complex that prevents us from keeping that mindset.
But when we embrace this mindset, we realize that it is applicable in many other areas than writing software. It is applicable to any goals that we try to accomplish in unordered domains, where the result of an action is literally unknowable. Internet startups are the perfect example for this: In a lot of cases it is he best option to ship something, and then respond to the market reaction. The crucial requirement for this fast-feedback cycle is to establish a "safe-fail" environment, where we can embrace failure and leverage failures to guide us to success.
Steve Freeman writes about Test-Driven Development and Embracing Failure and Obie Fernandez writes about Testing against business metrics that has nothing to do with code.
There is no measurement for technical dept. But the closest measurement is lines of code.
It is all very complicated but not magical!
Their "one-click-deploy" involved a telephone call, copy-paste sql-scripts and chanting & praying.
Software development is about creating something specific not something generic.
(Else you can send the customer a virtual machine, that´s generic.)
Take your last shipping project: Probably you could power-type all the code in one day … How do we spend our time?
About productivity: How good are you in thinking while typing?
If you have git, you have source control - if you have TFS, you have my condolences.
