[OSGeo-Discuss] Open Location Services

[Picavet Vincent]

Hi there,

> That said, the main theme of my enquiry still remains -- I 
> had never heard of "packaging" costs until now, and am 
> curious about quantifying them.

The term «packaging cost» is not really representative of the actual work it includes.
We are in fact speaking of code and project quality. In that sense, "packaging cost" are pareto's principle's 20% of a project. These 20% represent the work on the project which is never done for internal projects, as it takes a lot of time compared to what it's paying off. 

This may includes :
- technical documentation (as the project is internally developed, knowledge is available by asking individuals)
- end user documentation (sell training and expertise instead of writing non-paid end user documentation)
- code cleaning (while it's closed and intern, who cares about clean code ?)
- code documentation (ask your colleague if you want to know what «UGLY HACK» really means...)
- annoying little bug fixing (every team developer knows the non-documented workaround)
- setting up a dev environment (svn, ftp & stuff, managing rights, takes more effort for intern needs than an open dev env)
- setting up a project home website with basic information (never done for intern project if the product is not directly sold)
- setting up a support infrastructure and team (mailing list, specific persons in charge... done informally when intern)
- ...

If we were in an ideal world, all of this would be done for long even for small internal project. But in our real world, those points are often abandoned along the road of good intentions. It really depends on your internal project quality management. I've seen project with all the above done and well done, and other where none of it ever existed.

Successful Open Source projects however strongly need those points to be sure to gather a community of users, developers, and be able to reach a stable point where the project live by itself without perfusion. 

As to answer your specific questions :

> Imagine that I am a potential sponsor. You have developed <magic
> software> for your own company. A few users are expressing interest in
> that software. You write to the user list that you will put 
> that software into open source were your "packaging" costs 
> met. The following questions --
> 
> 1. How much are we talking about here?
> 
To stay with pareto's principle, i'd say around 20% of the initial project's price. It depends on the current quality of the code and project management type and infrastructur though. The closer it is from the «opensource way» of doing thing, the less these cost will be.
 
> 3. If no one comes up with the packaging costs, would you not 
> put it into open source, or would you still put it, but just 
> "dump the code into sourceforge" and let Darwin take care of it?

As far as a (my one at least) company is concerned, the idea is generally «open source quality software or don't», as bad code is bad image for the company. That said, there may sometimes be legal reason leading to open sourcing code without «packaging» it at all.

> 4. If you do put it in open source without any packaging 
> costs being paid to you, would you be losing out on any 
> particular revenue other than the time spent to put it into 
> open source?

Time spent to put it into open source is time not spent on other profit-making project. Then other revenue can be lost due to the opensourcing itself of the project, but this is another subject.

As to answer to miles about FLOSS project typology, I fully agree with him. There also is the kind of project written by an individual on its spare time, with high-standard quality code and infrastructures, which evolves, grows and turns into a successful big opensource project. Well, I'm still looking for examples, but I'm sure we can find some :)

vincent