[cvsnt] information currency in cvsnt? was: Re: [TortoiseCVS] information currency cvs in tortoisecvs distribution?

[J. Patrick Bedell]

Hi,
   Thanks for your reply, and for all your helpful work!
   Is there any possibility to incorporate information currency
issuance into cvsnt?
  The most functional example of information currency at this time is
with the server at http://prdownloads.sf.net/infoeng/icws-0.2.0.tar.gz
and the client at http://prdownloads.sf.net/infoeng/icsvn-0.0.3.tar.gz
.
   A public information currency service is available at
https://ic.infoeng.org:48443/icws/services/ICWS?wsdl , with an
interface at https://ic.infoeng.org:48443/icws that will eventually
(but certainly not right now :) be a stock-market type interface to
value the information that is used to create the information currency
that the running ICWS service issues.
   (I've posted the rest of my comments below.)

On 6/19/06, Arthur Barrett <arthur.barrett at march-hare.com> wrote:
> Patrick,
>
> > Is there any possibility to implement an information
> > currency interface, such as that within information
> > currency subversion (currently icsvn-0.0.3) into the
> > main line of Tortoise CVS development and release?
>
> >
> http://www.ietf.org/internet-drafts/draft-jpbedell-information-currency-
> 01.txt
> > which apparently describes it all.
>
>
> Sure.
>
> CVS, CVSNT and TortoiseCVS are all licensed under the GNU GPL which
> gives you the freedom to make these changes.  Remember GPL licensed
> software is "free as in freedom" not "free as in beer".  You can make
> any change you like and submit it back to the appropriate project for
> inclusion in the "canonical" code.
>
> If you are asking for *someone else* to implement it *for you*, then you
> have several options, that in summary are:
> * find someone else who wants this feature enough to implement it
> * pay someone to implement it for you (eg: March Hare Software who
> develop CVSNT)
>

   Only the information currency that effectively contributes to the
implementation goal will have value.  The purchase price of ic for
low-quality commits would be much less than that for ic from
higher-value commits.
   This is why issuance is based in the client in icsvn, the most
functional information currency client at this time.  Only the very
highest-quality (and economic value) commits would be incorporated
into the main line of development, based on, among other things, the
exchange value of the resulting client-issued information currency.
In other words, how high your selling price is for your information
currency after you generate it (with cvs commit's, icsvn ci, or some
other mechanism).

>
> > After a successful cvs commit, one or more information
> > currency series are generated for the code that was committed.
> > http://prdownloads.sourceforge.net/infoeng/ic-cvs-1.12.12.1.tar.gz
> > is an implementation of information currency cvs.
>
>
> Based on my limited understanding the change would initially go into
> CVSNT Server - probably as a plugin like the bugzilla integration or
> e-mail.  CVSNT Server has an API for creating 3rd party 3gl plugins on
> the server that can then intercept triggers and extend the behaviour of
> the server.
> The correct place to discuss this is the CVSNT newsgroup:
> http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
> or
> news://news.cvsnt.org/support.cvsnt
>
> Eventually Tortoise could be modified to pick up the tokens passed back
> in stdout to display you your "currency" (but I don't think this is the
> best way - see below).

   The tokens will enable people, eventually, to organize massive
amounts of information (in this case source code quality with iccvsnt)
in the form of markets.  The user interface for information currency
to manage software development is still under development, which is
why I started with tortoisecvs. :)

>
> I also assume that someone (you Patrick?) could already modify
> TortosieCVS based on the information that ic-cvs server returns (again -
> see my notes below).
>
> Note for anyone interested:
> * it looks like you need the "Information Currency Web Services"
> software installed first which requires "Java 5.0 (1.5) and the Ant
> build system" - all looks very unix-y/linux-y to me.
> * you then need to install the ic-cvs server (or ic-svn client)
> * interestingly the SVN version of this is implemented in the client -
> not the server, which seems less than intuitive to me.  Maybe SVN
> doesn't have any server side triggers?  Either way - implementing any
> payment system into something the "payee" controls (ie: the client
> software) appears unlikely to receive wide adoption.
> * if the goal of "information currency" is to pay a programmer based on
> what code/work they complete - then doing it "on commit" is probably

   The goal with icsvn client-side issuance to ensure that your
self-evaluation of your own code quality is as high as possible by
linking the exchange economic value of the corresponding ic to the
market estimation of your code quality.  This means that committing
hundreds of lines of comments at an appropriate point in development
will be extremely valuable an economic ic exchange, or actually
detrimental to your compensation at another point.  icsvn is the only
existing client right now (and not ic-cvs) that enables this
client-side estimation of commit quality, but development is
continuing to incorporate information currency features into cvs.

> incorrect - otherwise I'll just add hundreds of lines of comments and
> commit one line at a time to get my pay rise.  The currency needs to be
> issued on acceptance of those changes - so my proposal (should anyone be
> interested in reading this far) is to link this in with change sets (cvs
> commit -b).  When the change set get's "promoteed" (read: merged) onto a
> certain branch (eg: "accepted") then the information currency can be
> issued.  Presumably the "user id" would be linked into an RSA key (maybe
> via an extension to the CVSROOT/users file already used by viewcvs,
> e-mail plugin and bugzilla plugin).
> * Based on my above argument - when the currency is issued any response
> should be e-mailed to the submitter - so TortoiseCVS would in fact have
> no changes.
>
> I've cc'd anyone who I thought may be interested.
>
>
> Regards,
>
>
> Arthur Barrett
> Product Manager CVSNT
> March Hare Software
>