[Hdaps-devel] Re: [ltp] IBM HDAPS Someone interested?(Userspaceaccelerometer viewer)

[David A. Desrosiers]

 	(minor rant, in the future, please do not send the body of 
your message as an attachment, put it in the _body_ of the message 
instead. Perhaps your MUA is misconfigured, it makes it very difficult 
to reply).

On Wed, 13 Jul 2005 13:51, David A. Desrosiers wrote:

> You can trivially (as you are a programmer) put a filter in *your* 
> system that reformats code, patches, etc. any way you want.  You 
> don't need to ask other people to do anything to accommodate your 
> style.

 	And that's precisely what I do, with indent(1) here, when 
patches or code is sent to me in-full, not as diffs in email.

> Have your computer format code the way you like it - don't tell 
> other people to format their code for you.  It's simply not 
> important enough to bother even writing to them about.

 	When you get 20+ patches from people per-day, with potentially 
hundreds of lines of fixes or changes, and those diffs are in some 
random, unfamiliar indenting format (and trust me, I've seen some 
WACKY ones), you can't just reformat the diff while it remains an 
attachment to the message.

 	You have to detatch it, save it to a file, run indent on it 
with your specified parameters, find/fix/comment on it, then reply 
back to the original person with any specific updates to their diff.

 	If the diff matched your coding style, you could immediately 
see or spot any issues right away, and comment on it without having to 
take the extra 4 steps. Multiply that out by 20 diffs per-day and you 
can see that it really cuts down on the time required to respond to 
patches and requests if you receive the code in the same way that the 
users received the code.

 	And that doesn't even count the users who do the same thing. 
I've had users take a 65k source file, reindent the whole thing, patch 
a 1 or 2 lines, then send it back to me as a unified diff. Do you know 
how difficult it is to have to find those 2 changed lines across that 
130k diff file?

> It's your computer - use it.  My comment about the compiler not 
> caring was to point at the fact that the information does not change 
> depending on re-formatting.  So reformat it, by all means.

 	Actually that's not entirely true, for example:

 	if (foo)
 		bar;

 	vs.

 	if (foo) {
 		bar;
 	}

 	Can have very different results, depending on what follows 
below that construct. In general cases, you might be correct, but 
changing someone else's code formatting is generally not a good idea, 
regardless of whether or not tools exist to reformat it, those don't 
model every situation where real-life enters the picture.

> > My project, my rules as they say.

> Your choice - your responsibility.  Stop laying it on other people. 
> How's that?

 	Its easy to stop laying it on other people. They can simply 
have their patches/code rejected, and they can find another project to 
work with, if they choose not to deliver things in a way that is 
expected. What if I received an update to my project, and all the 
comments were in Russian? Should I be expected to go find a Russian 
translator to decipher them for me, and then convert them to English 
before I commit and test the code? No. When in Rome...

> Is it clearer this time?  Telling other people to accommodate your 
> personal choices that you could handle with about 0.0000003 cents of 
> electricity is ridiculous.

 	It isn't about saving cents on electricity, its about working 
in a way that is most productive for the person responsible for that 
work, in this case the developer. See above.

> Sorry to bludgeon you with it, but your brain armor seems kinda 
> thick there.

 	If you'd like to throw around insults.. your lack of 
understanding how developers work most efficiently shows a clear use 
of very narrow vision in your thought process.

> And I'm a Software Quality Assurance Analyst for a living, mate, for 
> the last 15 years.  Yes, I know how to write bugs.  But what I said 
> isn't about bugs, it's about asking other people to take care of 
> something your computer can do in the time it takes me to type the 
> 'e' at the end of 'sentence' - format the code any way you like.

 	In your SQA world, that may be true, but in development, that 
isn't always true. When you receive dozens of different (sometimes 
overlapping and conflicting) patches for code per-day, its VERY time 
consuming to have to reformat each one after detaching it from the 
mail envelope just to see if its a viable patch. Many patches are 
rejected outright before they're even applied based on visual review 
of the contents. Reformatting that code before patching and sending a 
diff can have dramatically different results on how that code is 
viewed.

> I'm sorry, but I'm close to falling over laughing - this is 
> so...Middle Ages, so... Kafkesque.  You're a programmer, right? 
> Then, you must have a computer...right?  Can't it format code to 
> your heart's content?

 	Please continue laughing, get it out of your system, and when 
you're done, talk to some real developers about how they prefer their 
code to be constructed. Perhaps you've been in SQA too long and you 
don't see how other people function. Just because it works for you 
doesn't mean it works for everyone else.

> This isn't about bugs.  It's about misplaced world views.

 	No, its about catering to the developer's wishes. If you want 
the developer to be responsive to your patches and requests, you 
provide the information in the best way possible for him/her to act 
upon them. The harder you make it for the developer to continue to do 
his job, the longer it will take for action on that issue.

 	Following your logic, why not just submit all patches as an 
XML file then? I mean this is 2005, right? Get with the times and all.

> Sorry, mate, I've been chuckling over this one all day.  Should we 
> all sit in a particular posture or wear special clothes when we 
> write to you, too?

 	When you're in my project, working with my code, you can 
either format the code in the way I specify, or go elsewhere. That's 
your choice, and you are free to make it. I don't care about your 
clothes or your posture, since that doesn't affect me. Once you start 
telling me what *I* should wear, then we're going to have a problem.

> Sorry.  I tried to be Politically Correct and sensitive and 
> diplomatic with my first message.  I guess I'm just not any good at 
> diplomacy...

 	You might do well trying to convince the thousands of Linux 
kernel developers to change the way they operate too, by taking 
patches in any old format they receive them in, and forcing the 
developer to do the work of deciphering the innards, reformatting the 
code back the way they expect it to be, etc. What if code updates were 
sent in Word documents, postscript files, XML files, different 
languages and other dreck? Would you expect the developer to go 
install Windows, cut and paste out of postscript, use expat/parsers to 
dump the data from XML, etc.? No, you certainly would not.

 	In any case, its your opinion, and I'm free to ignore it.



David A. Desrosiers
desrod@gnu-designs.com
http://gnu-designs.com