[lug] Development & reporting tool choices

[Paul E Condon]

On 2009-02-05_19:25:52, Paul E Condon wrote:
> On 2009-02-05_18:34:16, David Morris wrote:
> > On Thu, Feb 5, 2009 at 16:50, Kevin Kempter <kevin at kevinkempterllc.com> wrote:
> > > Hi All;
> > >
> > > We're preparing to embark on a new development project. It will eventually be
> > > released dual projects similar to Red Hat with a commercially supported side
> > > and a leading edge open source side.
> > >
> > >
> > > The project will pull data from a database and present info to users. We'll
> > > need a graph/chart component and a way to generate html, PDF and spreadsheet
> > > reports/downloads. It will be a web based interface.
> > >
> > > Questions:
> > >
> > > 1) I'm thinking we'll write the initial commercial (closed source) code in C.
> > > Anyone have any thoughts/alternative suggestions ?
> > >
> > > 2) anyone know of a good source of C library routines (something like CPAN) ?
> > >
> > > 3) Any thoughts on the methods to generate XML via C ?
> > >
> > > 4) Any suggestions per the generation of the PDF, HTML, etc reports on screen
> > > and for download ? maybe php within the web front end? a set of C routines to
> > > be called based on basic report parameters? others ?
> > 
> > Not certain if the language is flexible, but unless performance is an
> > issue, I might be tempted to use Python instead of C.  The Python
> > language has some fantastic libraries for database access, XML
> > handling, string parsing, PDF generation, and web-page interfacing.  A
> > lot of functionality comes in the standard libraries shipped with the
> > language, the rest in widely used third-party libraries (such as
> > ReportLab for PDF generation).  You can also easily interface to C/C++
> > code if needed using SWIG.
> > 
> > It is still possible to provide closed-source binaries for python
> > code, FYI.  I haven't done this myself, but I know there are a couple
> > of different ways to handle it.
> > 
> 
> Choosing to go with the closed source development first seems to me putting
> the cart before the horse. There are many open source library packages, and
> gcc is a quite adequate C compiler. Using open source tools to produce 
> software does not commit you to always being rabidly open source. Using free
> (as in beer) tools will make your project much less costly. Having your early,
> buggy code open to help from interested computer people will be good for the
> project. ... all the usual arguments why open source is good ... 
> 
> just my $.02

Sorry David, I meant to address the comments to Kevin

-- 
Paul E Condon           
pecondon at mesanetworks.net