Joel Spolsky has just published another interesting blog entry called Evidence Based Scheduling. As so often in recent times the blog is about the latest release of FogBugz and some of its new features. It is hard to restrain the feeling that Joel is using his blog more and more as a marketing platform, but on the other hand that's probably what it is and who wouldn't knowing how many people one can reach.
Anyways, as the name suggest Evidence Based Scheduling it is about scheduling software and at its core it also deals with estimates of individuals. You probably think now:"Oh god, another one!". In fact the ideas behind Evidence Based Scheduling are not new. A little bit of SCRUM, a little bit of probability theory and voila you have Evidence Based Scheduling. What is new is that there is (supposedly) tooling around it. A lot of books tell you how to do it in theory, but none gives you the tools to do it. And with tools I mean an easy means to gather the required data to make the framework work - and in case you are wondering, excel spreadsheets are not cutting it!
To be honest i haven't tried FogBugz yet, but this new feature might get me to do it. Maybe it really solves one of the big problems of estimation frameworks - gather the relevant data! I would love to hear if someone has already some experience with FogBugz. Anyone? If not, make sure to drop by soon again. There will be more on this shortly.
--hardy
Monday, October 29, 2007
Monday, October 22, 2007
Huge web design survey
Awesome blog A List Apart has published the results of a web design survey answered by close to 33,000 web designers. Lots of demographics has been gathered to get an overview of the field.
Snapshots: the table showing salary range by longevity field reveals that 59,8% of those with less then one year of experience makes less then $20K while 54,1% of those with 9-10 years experience makes more than $60K. Folks working for the government are less likely to quit. Project managers and information architects are most satisfied with what they are doing (>50%) while web masters are least satisfied (40,3%). And like this it goes on and on. One interesting fact is that a salary between $40K and $60K is the most satisfying. It is as satisfying to earn more then $100K as it is to earn between $20K and $40K.
An interesting report with interesting conclusions. A must read for any recruiter in the field. The report is available for free: The attached report shares everything we learned. We offer it freely to this community that has given us so much.
Snapshots: the table showing salary range by longevity field reveals that 59,8% of those with less then one year of experience makes less then $20K while 54,1% of those with 9-10 years experience makes more than $60K. Folks working for the government are less likely to quit. Project managers and information architects are most satisfied with what they are doing (>50%) while web masters are least satisfied (40,3%). And like this it goes on and on. One interesting fact is that a salary between $40K and $60K is the most satisfying. It is as satisfying to earn more then $100K as it is to earn between $20K and $40K.
An interesting report with interesting conclusions. A must read for any recruiter in the field. The report is available for free: The attached report shares everything we learned. We offer it freely to this community that has given us so much.
Sunday, October 07, 2007
Order does matter
Ever used profiles in maven? Great stuff. Ever tried to trigger multiple profiles? Great in theory, but unfortunately in this case maven is not delivering what one would expect - especially when you have the the same properties in multiple profiles and you want the values of the 'last' profile activated to be the effective ones. This is exactly how settings.xml works. If a profile in settings.xml is activated and it defines a property which was already set by another profile it overrides its value. Not so with multiple profiles defined in pom.xml. Say we have to two profiles default and override defined as follows in pom.xml:
<profile>
<id>default</id>
<properties>
<test>DEFAULT</test>
</properties>
</profile>
<profile>
<id>override</id>
<properties>
<test>OVERRIDE</test>
</properties>
</profile>
Running the build via:
> mvn install -P default,override
one would expect that the value for property test is 'OVERRIDE', but it is not :( It turns out there is a bug in maven which causes the order of the profile activation to be undetermined. The good news is that this particular problem is solved - in version 2.1-alpha! This means until maven 2.1 is available watch out when trying to use profile overriding. It can save you a few hours of tearing at your hair trying to figure out why the build is not behaving the way you want it to behave.
<profile>
<id>default</id>
<properties>
<test>DEFAULT</test>
</properties>
</profile>
<profile>
<id>override</id>
<properties>
<test>OVERRIDE</test>
</properties>
</profile>
Running the build via:
> mvn install -P default,override
one would expect that the value for property test is 'OVERRIDE', but it is not :( It turns out there is a bug in maven which causes the order of the profile activation to be undetermined. The good news is that this particular problem is solved - in version 2.1-alpha! This means until maven 2.1 is available watch out when trying to use profile overriding. It can save you a few hours of tearing at your hair trying to figure out why the build is not behaving the way you want it to behave.
Monday, October 01, 2007
Chad Fowler on the Big Rewrite
Interesting stuff from Chad Fowler on why doing the Big Rewrite may be a bad idea. My current customer has failed all attempts at doing Big Rewrites - actually not a single old system has been replaced in the last 10 years despite numerous attempts.
Subscribe to:
Posts (Atom)
