nALFS programming details

James Robertson jwrober at linuxfromscratch.org
Tue Aug 10 18:25:02 PDT 2004


Kevin P. Fleming wrote:
> James Robertson wrote:
> 
>> I have seen too many development projects go to crap because the team 
>> did not think the whole pie out before hacking out code.  If we get 
>> the whole set of features/requirements down on [virtual] paper with 
>> details as to what we are thinking about, this will really help us 
>> down the road.
> 
> This is true, but we also have to realize that since this is a volunteer 
> effort there is only so far we will be able to take the analysis before 
> it's time to start building testcases and demonstration programs.
> 
> We don't have the luxury (burden) of working as part of a contracted 
> development team where we must follow a rigid path from requirements 
> analysis to implementation... that's both good and bad.

I agree, but I still think we need to noodle this a bit more before 
selection of programming language, libraries, etc, for the real "go". 
Demos and testbeds can be done in anything like you said and can start 
now AFAIAC.  I am all for that.

>> With this information in hand, we can come up with a clear roadmap as 
>> to what happens when, what programming language will be used, 
>> toolsets/libraries selected to fit our needs.  We can use scripting 
>> languages (like what Kevin said) to build quickie proof of concept 
>> stuff to help flesh out our ideas, but those would happen during the 
>> requirements definition phase.  Some scripting languages could be used 
>> later for different clients, but that is not really an issue - what is 
>> an issue is having a clearly defined direction that we all agree on 
>> before writing any real code.
> 
> 
> Agreed.

Kewl
James



More information about the alfs-discuss mailing list