Should I write code before I can test it?

Maybe this would be better as a question on twitter, as I may not get to many comment here on my web 1.5 journal.

Is it OK to write code I cannot test and assumes that by the time I am done writing the features I have in mind, the environment to test it would be available, or should I start setting up the environment first, and get the hello world application up and running with the monitoring and clean log files, so that I can then go ahead with small incremental cycles which are all providing features I can test as I go?

On another topic, when can you consider that a forum or forum category is to be considered dead?
a) 0 post in a day
b) 0 post in a week ?
c) 0 post in a month?

Comments

Kati Molnar said…
I did both in the past. There is nothing like the incremental coding when it comes to understand the project better.
When time allows, I find it helpful to leave some of the bottom up pieces as introductory toys and do a full redesign from the top eventually.
The situation you have been in April looks all too familiar to me. I am just waiting to get access to the environment while I have a pressing deadline. So one starts bottom up, doing the little parts that are needed, easily understood, and eventually can be fitted into the big picture.

Hey, programming is fun, we can enjoy it as well (ahh, if I could forget the deadline looming...).

Kati
10:07
Post a Comment

Popular posts from this blog

Changing the version of JDK used by JDeveloper

I have recently learned a quick and lazy way to change the version of the JDK used to run JDeveloper 10g R3. I am sure there is an option somewhere in the configuration screen to do the same, which I couldn't find... While JDeveloper is not running, locate the embedded JDK on disk and rename (or delete) that directory. For me, it was C:\opt\jdev\jdk. Once it's done, restart and you will by asked to locate the version of the JDK you want to use. It's another step to get going with JAVA on Vista. -ecco
Read more

Connection reset from a WCF Web Service

Here is an error message you may see, when calling WCF Web Services from java: IOException retrieving the response: java.net.SocketException: Connection reset As it may not be obvious the first time you see it, this may save you some time. One reason for this behaviour is when the WCF endpint is expecting to see wsa:Action SOAP header in the request. To find out if you are facing this case, you have to look inside the WSDL, and search for the policy that describes the contract for your endpoint. <wsp:Policy wsu:Id="CustomBinding_Echo1_policy">   <wsp:ExactlyOne>     <wsp:All>       <wsaw:UsingAddressing/>     </wsp:All>   </wsp:ExactlyOne> </wsp:Policy> To make this a concrete example, here is a sample WSDL . To find the policy, you will need to follow the import as the policy are not always advertized in the top-level document. Here is how the import looks:   <wsdl:import namespace="http://example.org/" location="h
Read more

unexpected null value for literal data

There is a challenge that is hard to overcome at first; making sense of error messages. You can always Google with the text of the error, “unexpected null value for literal data” and hope to be lucky . If you are using Oracle Application Server 10.1.3.0, here is the explaination for the message and two ways out. The Exception: unexpected null value for literal data at oracle.j2ee.ws.common.util.exception.JAXRPCExceptionBase.<init>(JAXRPCExceptionBase.java:52) at oracle.j2ee.ws.common.encoding.SerializationException.<init>(SerializationException.java:26) ... The relevant schema: <complexType name="Customer"> <sequence> <element name="address" type="tns:Address"/> <element name="firstName" type="string"/> <element name="lastName" type="string"/> </sequence> </complexType> A sample code that will cause this exception: public void testServ() throw
Read more