Observer Effect

Scientists have to deal with the observer effect, which means that observing something actually changes it. Typically we think of quantum physics, where this effect is very strong and surprising and closely related to the Heisenberg Uncertainty Principle, but it is actually something that in a more abstract sense is present in a multitude of situtations. Just think of human interaction. If we want to find out about people, we can ask them. But this conversation actually changes the people, sometimes in a way that we can neglect or tolerate.

But we also have this in the case of IT. If we think of a software and we want to observe if the software behaves well, we need ways to observe the software. Very often we use logging, sometimes monitoring tools, and sometimes debugging or even profiling. We think that they do not hurt us, apart from using resources, but we have to be quite careful. The example of logging is quite good, because it is quite common and usually something that we do a lot, without wasting too much thoughts about it.

Now logging slows our application down that is known already. Now we tend to use a slightly less noisy log level, because terabytes of log are still a pain, even today. But usually the messages are calculated and then discarded by the logging framework. With functional language features there are quite elegant ways to deal with this, by just passing a function that calculates the message on demand instead of passing the message. It has always been possible, but too clumsy to actually do it, unless the logging framework can rely on macro facilities, even such simple ones as the C-preprocessor. The deferred evaluation has its dangers as well. If an object that is passed as an ingrediant for a potential message already changes, while the message is created, we might get funny effects. Maybe only in the log, but maybe it could crash the application or stop the main program flow from doing its work. We need to be careful, unless the object is immutable.

In case of Hibernate or JPA or similar frameworks this can be specially interesting, even with eager message calculation. Accessing attributes of the object can actually lead to database operations. They can fail. They can create load, maybe deadlocks. They can have lost their transaction. A lot of things can happen in places far away from where we assumed to do the DB-work. This actually changes the objects. Do we want such operations to occur during logging? Maybe differently depending on the log level? Immutability is our friend, especially in conjunction with JPA, but that is a long story. We may at least be lucky that we actually have some tables that we only read. We can make the objects „pseudo-immutable“, but still JPA-magic must mutate them at least during the read operation. It is tempting to let tools generate the toString-methods of objects, but it is very dangerous here. We should avoid including any potentially lazily loaded attributes in the toString-output, because otherwise they will be loaded during the logging or even worse differently depending on how we log.

The next thing is the NullPointerException during logging. It is quite common in Java, for example. And we do not want to burden our program logic with NullPointerExceptions from the logging, especially not with those that occur only sporadically. So it is a good idea to be careful and to test well. Only the combination is possibly good enough.

Modern times create more demand for some kind of real multi threading, not in the JEE-sense with a couple of EJBs that can run in parallel, but with massively parallel operations. Even though we have a multitude of logging frameworks and unifying logging frameworks and even more of them, there is a common weakness that they tend to share. Writing into one single target is achieved by some kind of synchronizing, which can slow our application down and change the timing behavior in ways that we did not desire. Asynchronous logging could be good, but in a way this only shifts the problem a bit.

Share Button

Functional Programming: Article in another Blog

I wrote a Guest Blog Article about Functional Programming on Adesso’s Blog in German.

Share Button

Testing and Micro Components

It is an interesting trend to create micro components, micro services and use building blocks that are very easy to understand. As always, a new approach will not eliminate all problems. The problem that we are solving has an inherent complexity that we cannot beat. But usually we create a lot of complexity that would not be necessary to solve the problem and this can be avoided. Any approach that really helps us here is welcome. If and how micro components and micro services can help us, is another issue, I will just assume that they are in use.

Think of constructing a building. Now this is a very complex task, but we simplified it by building Lego pieces instead. They are simple, easy to understand, well tested and we just have to compose them to get the building. The building does not become trivial buy that, but maybe it becomes easier to create it.

In the end we want to sell a building. The fact that it is constructed of ISO and TÜV and CE and whatever certified lego blocks is not really relevant for the users of the building. They actually want a building, not lego blocks. Maybe they do not even care how it has been made, as long as it is good. They should not care, but usually they do at least a bit. We do not live in an ideal world and I would care too.

Now the building has to fit into the city. It has a function to play. It is itself a lego block for building the city. But for the moment we are actually only doing one building, not a whole city. So we need to provide tests that it works with its environment and provides the functionality that we expect. There is an API and an API-contract. We need to test this API in the way it will be used in production. That is on Linux, if the productive servers run Linux, not on MS-Windows. But Linux is anyway better for developement than MS-Windows, unless we are working on MS-specific technologies, where the servers will run MS-Windows anyway. We have to test with the database product that is used in production, for example PostgreSQL or Oracle. Tests with In-Memory-DBs or products like sqlite are interesting and maybe helpful for finding bugs by having tests that run presumably faster or with less overhead, but there are ways to provide the real DB to each developer and the overhead is not much, once this has been prepared. And a local Oracle instance is not much slower than an in-memory-DB, but allows us to inspect and change data with SQL, which can be very useful. And if we use application servers or middleware, it is mandatory to test with the same middleware as in production, no matter how much compatibility between the implementations of different vendors has been promised. All of these issues are in theory not existent, but I have seen them in practice and that is what counts. Yes, I love theory. 🙂

The other thing is that the fact that the lego pieces have been tested does not guarantee that the whole building will work. The composition creates something special, that is more than the sum of its pieces. That is what we want to achieve. So we need to test it.

So it is good to have automated tests that run against the service using the API that is used in production, possibly REST or SOAP, for example. And to run the test against an instance of the application that runs on the same kind of server against the same kind of database and on the same kind of middleware as in production. At least continuous integration should work like that. The smaller our components are, the more important it becomes to test them in conjunction. Problems will occur there, even though the components by themselves seem to work perfectly.

Now there is demand for running more local tests, unit tests, as they are commonly called, to actually test lego blocks or small groups of lego blocks. It is good to locate problems, because the end-to-end-tests just provide the information that there is a bug, but it is hard to locate. And without thorough testing of the sub components it is possible that we overlook edge cases on the overall tests that will anyway hit us in production, because only the more local tests can explore the boundaries that are visible to them, but obfuscated when accessed indirectly. Also we do not want to be happy with two bugs that cancel each other, because they will not always work together in our favor. So in the end of the day it is important to have automated tests at all levels. And to make them an important issue.

I have seen many projects, where Unit-Tests eroded. They were once done, but did not work any more and were impossible or hard to fix, off course lacking the time to do so. Or the software was hiding access points that were needed for testing. Just two examples for this, for our Java friends: methods can be package private and test classes in the same package to allow testing. And application servers need to actually allow certain remote access, usually via REST or SOAP these days. And we need remote debugging access. This is all easy to achieve, but can be made hard by company policies, even on development workstations and continuous integration servers.

Share Button