Comments for TuxTips http://www.tuxtips.org Linux Tips and Screencasts Wed, 03 Jun 2009 00:53:21 +0000 http://wordpress.org/?v=2.9.2 hourly 1 Comment on Lines of Code – The Most Meaningless Metric by admin http://www.tuxtips.org/?p=4&cpage=1#comment-9 admin Wed, 03 Jun 2009 00:53:21 +0000 http://www.tuxtips.org/?p=4#comment-9 asharif: Thanks for taking the time to comment. The number of lines of code may be a useful piece of data when calculating more complex code metrics, but by itself it still says little about the code. The reason LoC is a useless measurement (or worse, one that has negative value) is because you need other data to place it in context, but for some reason people still act as though having this single number tells you everything you need to know about the code. asharif:
Thanks for taking the time to comment. The number of lines of code may be a useful piece of data when calculating more complex code metrics, but by itself it still says little about the code. The reason LoC is a useless measurement (or worse, one that has negative value) is because you need other data to place it in context, but for some reason people still act as though having this single number tells you everything you need to know about the code.

]]> Comment on Lines of Code – The Most Meaningless Metric by asharif http://www.tuxtips.org/?p=4&cpage=1#comment-8 asharif Tue, 02 Jun 2009 16:45:07 +0000 http://www.tuxtips.org/?p=4#comment-8 I think the author has been mislead a little bit on the LOC metric. LOC, as I understand (with a BSCS and a MSSE) is not a measure of complexity or effort. It's a metric that helps calculate other metrics that measure complexity or effort. For example a software component could have 100 KLOC and a cyclomatic complexity of 200 and another software component could have 100 KLOC and have a cyclomatic complexity of 20k. Here the LOC is constant yet the cyclomatic complexity (which is calculated with LOC as a factor) is not. As far as the $ amount assigned to LOC, well there is a good reason for that as well. Again, it's not LOC by it's lonesome. Formulas such as COCOMO II use LOC as a factor in the formula. There are other metrics that have to be in place before you can even try to use LOC for cost. The primary one being how many LOC does a team produce per month on average. this average has to come from a good sample set and cannot just be made out of thin air. Once you have this LOC per month number then it's perfectly rational to plug it into a cost calculating formula such as COCOMO II. Functional point analysis is very similar, but it removes LOC and replaced it with functional points. FP being how many FP can be accomplished per month. Then you plug that FP in the formula to get an effort estimate. Bottom line here is that LOC is a metric used in software engineering. It's a valid metric when used in the correct context. It's like using an incorrect constant in any other equation in math or physics...it will net the wrong answer. I think the author has been mislead a little bit on the LOC metric. LOC, as I understand (with a BSCS and a MSSE) is not a measure of complexity or effort. It’s a metric that helps calculate other metrics that measure complexity or effort. For example a software component could have 100 KLOC and a cyclomatic complexity of 200 and another software component could have 100 KLOC and have a cyclomatic complexity of 20k. Here the LOC is constant yet the cyclomatic complexity (which is calculated with LOC as a factor) is not.

As far as the $ amount assigned to LOC, well there is a good reason for that as well. Again, it’s not LOC by it’s lonesome. Formulas such as COCOMO II use LOC as a factor in the formula. There are other metrics that have to be in place before you can even try to use LOC for cost. The primary one being how many LOC does a team produce per month on average. this average has to come from a good sample set and cannot just be made out of thin air. Once you have this LOC per month number then it’s perfectly rational to plug it into a cost calculating formula such as COCOMO II. Functional point analysis is very similar, but it removes LOC and replaced it with functional points. FP being how many FP can be accomplished per month. Then you plug that FP in the formula to get an effort estimate.

Bottom line here is that LOC is a metric used in software engineering. It’s a valid metric when used in the correct context. It’s like using an incorrect constant in any other equation in math or physics…it will net the wrong answer.

]]> Comment on Journey to Python Part 1: First Steps by codeninja http://www.tuxtips.org/?p=13&cpage=1#comment-5 codeninja Sat, 16 Aug 2008 20:07:37 +0000 http://www.tuxtips.org/?p=13#comment-5 noor: Thanks for the response; I'm sure that typing will come up pretty soon. noor:
Thanks for the response; I’m sure that typing will come up pretty soon.

]]> Comment on Journey to Python Part 1: First Steps by noor420 http://www.tuxtips.org/?p=13&cpage=1#comment-4 noor420 Sat, 16 Aug 2008 13:06:37 +0000 http://www.tuxtips.org/?p=13#comment-4 Great website. I am new to python also and am reading this series of articles with lot of interest. Also, if you don't mine, can you tell us how you feel about the the static vs dynamic typing (c++ vs python in this matter). Thanks a lot. Great website. I am new to python also and am reading this series of articles with lot of interest.

Also, if you don’t mine, can you tell us how you feel about the the static vs dynamic typing (c++ vs python in this matter).

Thanks a lot.

]]> Comment on Journey to Python: Introduction by qui http://www.tuxtips.org/?p=12&cpage=1#comment-3 qui Tue, 12 Aug 2008 12:40:01 +0000 http://www.tuxtips.org/?p=12#comment-3 I am so glad I have discovered this, last week I also decided to start learning Python (check my site, I'm not lieing!). So perhaps this will be a journey for the both of us eh? I am so glad I have discovered this, last week I also decided to start learning Python (check my site, I’m not lieing!).

So perhaps this will be a journey for the both of us eh?

]]>