[kwlug-disc] the intertoobz never forget

[unsolicited]

Chris Frey wrote, On 10/10/2009 4:17 AM:
> On Sat, Oct 10, 2009 at 02:17:53AM -0400, unsolicited wrote:
>> I'm mostly thinking about one-off app development here, for internal 
>> use, rather than an on-going development effort for an evolving app 
>> that you expect to use and grow with for some significant period of 
>> time. Is the latter the direction from which you're coming on this?
> 
> I'm referring to John's point:
> 
>>   Not quite. I meant the more people and companies that rely on Open source
>>   the more contributions will be made to support it the software they rely
>>   on for income.  Not only will small consultants provide changes to
>>   specific applications (Drupal??, etc), but core technology vendors like
>>   Intel, AMD, IBM and HP will help to advance core technologies like Linux.
>>
>>   Changes that are made by end users do not have to be distributed.
> 
> so...
> 
> On Sat, Oct 10, 2009 at 02:17:53AM -0400, unsolicited wrote:
>> Chris Frey wrote, On 10/10/2009 12:02 AM:
>>> And always going back to the source code to include your own changes
>>> eliminates some of the advantage of using a distribution.
>> But this last implies widespread use of your modifications within your 
>> organization. I would have expected such changes to be more one off.
> 
> An application written in-house (whether one-off or not) is not likely
> to be contributed back, in my opinion.  It would need to start its
> own project just for that app, and if its logic is focused on internal
> company behaviour, then the only benefit the company got from writing it
> was using an open platform.  There's little use in sharing it, unless it
> really is general use, or you're a Netscape that wants to donate Mozilla.
> 
> I'm referring to the tweaking of, and adding features to, an existing
> open source project -- projects that actually have an upstream to send
> patches to.
.
.
.

Well put. I had lost that we were talking about the modification of an 
existing project (which by definition then is not one-off, and not 
company specific). I was thinking more of a unique or initial effort 
being put out by an organization.

Thus, your point, that if it's so unique to a particular company it 
wouldn't be set up as a project in the debian (say) repository.

But I would encourage it to be put up. Many things really aren't as 
unique to ourselves as we think, and maybe someone else out there will 
participate and start to gain you all the advantages of being a public 
project. Bug fixes, testing, enhancements, etc.