For people writing drafts.... |
||
 Ret. Gen D-Cypell Guys... firstly... excellent work on the draft algorithms here. Great to see some real gray matter churn on this board :o). I love this stuff!! :o) A point though... more important than posting up implementations of algorithms is posting up example inputs and expected outputs. The actually bit-shifting work can be done in many many many ways and in different languages, but what we really need is something we can use to TEST if the algorithm is working correctly. For example... L33tSpeak --------------------- Input: D-Cypell Rules Output: D-Cyp3ll Rul3z Provide a few of them before posting your algorithm. Then someone can come back and say ahhh... but for THIS input... <input here> I would expect this output <output here> but what I get from your code is <actually output>. When this occurs, we need to either correct the code or correct the expectation. What we need is as many example inputs and outputs as possible. When it comes to actually implementing the software we can write unit tests that will throw each example input at the algorithm and check that we got the expected output. This can be built into the build process so that every change is tested. If someone tweaks an algorithm for performance reasons they might break something and not realize, but thats OK... the tests will fail and we will know its broken. Hopefully I have adequetely express why the input output examples are more useful than the code. The code will change as things evolve but the input/output will remain far more static. Also it means that when we are discussing implementations we are all working with the same expectations. On 2005-02-15 17:19:31, snarkles wrote >his thread is to discuss the development of the desktop version of SNEAK. Replies:
|
||
| CyberArmy::Forum v0.6 Generated In 0.02262 seconds |
2005-02-15 17:19:31
2006-01-03 05:04:05
2005-03-21 03:08:22
2005-03-14 19:01:39
2005-03-24 22:41:05
2005-02-17 20:09:14
2005-02-15 20:12:52