I have to register with Andrea under the dissenting opinion here. The
"invent here" policy does not sound like a good idea to me. In fact,
it seems to eliminate one of the great benefits of the Java language,
widespread use of popular and task-specific programming languages. If
you want to use a "every project roll their own solution" philosphy,
start programming in C. :]

Seriously folks, there is no way I'm going to spend two weeks, or even
a week, working on some code that already comes packaged in a JAR that
I can download and add to my classpath in about 30 seconds.

Not only that, but as soon as you roll your own solution you
eiliminate a team of developers with expertise in a certian aspect of
the Java rogramming language. Take Joda as an example, how many
programmers do we have that now Time/Calendars like the Joda guys? How
much time can you sock to Time/Calendar programming challenges. The
Joda guys do it every day...

I think we are attempting to treat a problem with the wrong medicine.
If dependencies are an issue in GeoTools let's look at a better system
for organizing dependencies. Can I see a list of the dependencies of
each module and the JAR version within a couple mouse clicks? How do
we decide as a group to move to a new version of a library? Can we
package the libraries used by the majority of our modules into a
convenient download for our users?

I think the policy should be to avoid libraries that provide duplicate
functionality, not to avoid use of libraries. For example, it seems
silly

Landon

On Mon, Jun 30, 2008 at 7:10 AM, Andrea Aime <[EMAIL PROTECTED]> wrote:
> Jody Garnett ha scritto:
>> I was looking at the "how to import a shapefile" page on the wiki - and
>> it mentions the "do not invent here" policy and how that has lead to our
>> massive amount of dependencies. Here are some ideas on how to set up a
>> "positive" invent-here policy.
>>
>> How about this for a GeoTools 3 where live is beautiful etc...
>>
>> GeoTools 3 should adopt an "Invent here" policy:
>> 1) unless a dependency offers significant benefit we should roll our own
>> - A dependency that brings in additional dependencies is a bad thing
>> - A dependency that is used by several modules is a good thing
>> 2 Significant benefit is measured in weeks not days
>> - if you want to use Joda time it better offer more functionality than
>> we can expect you to code in in two weeks. If you are using a couple of
>> functions we can probably expect you to be able to code them up in 13
>> days... you do after all have Joda time to serve as an example.
>
> Ideally this looks like a very good idea.
>
> In practice I don't know how I can justify spending 2 weeks of
> resources when the same work can be done in 2 hours.
>
> Turning two weeks into a monetary evaluation, even assuming a
> low rate of 500USD per day, that's 5000$ (assuming only 10 days,
> the working days in two weeks)... can you justify
> that much with your customers to avoid a dependency?
>
> Or if you want to turn that into a pure open source metric,
> let's assume someone that, like me, was working only in
> his spare time after work and in the weekends, someone
> being able to work the equivalent of 2 days per week
> on gt2. That turns the effort into an effective 5 weeks.
> Do you believe this is reasonable, especially for someone
> that's coding "for fun"? I don't think so?
>
> Cheers
> Andrea
>
> -------------------------------------------------------------------------
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services for
> just about anything Open Source.
> http://sourceforge.net/services/buy/index.php
> _______________________________________________
> Geotools-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/geotools-devel
>

-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
Geotools-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geotools-devel

Reply via email to