It sounds like there are a lot of good ideas for reading file formats and alike. This 
is great, and commons is a natural home for it.

However, what must be considered is here is community. Starting a new project is 
difficult as it involves gathering a group of people to work on the code, and maintain 
it. [math] is an unusual recent example in that it has successfully built a new 
community. Most new ideas fail because of community.

For this csv and other reading code I believe that it would be better to build around 
[io] or [codec]. There is already some code and some community to start from. It also 
limits the number of dependencies/new projects.

Stephen

>  from:    Simon Kitching <[EMAIL PROTECTED]>
>  date:    Thu, 26 Jun 2003 08:04:51
>  to:      [EMAIL PROTECTED]
>  subject: Re: [SURVEY] Commons-csv or not?
> 
> Hmm .. an "import-stuff-into-xml" project? Interesting...
> 
> I have in fact written exactly this for my current employer, for a
> (continuously expanding) series of formats. We then apply stylesheets to
> the results, for various purposes.
> 
> I doubt I could contribute any code, but can definitely confirm that I
> would have used such a library if it had existed about a year ago when I
> started (and had been sufficiently flexible/complete)!
> 
> I could also contribute a collection of weird and wonderful formats as
> invented by our clients :-)
> 
> Here's a few possible formats in addition to csv:
> * ASN.1
> * windows .ini files
> * java .properties files
> * EDIFACT
> * apache-httpd .conf format :-)
> 
> And here's some potential "competitors", which also import data into xml
> formats:
> * Microsoft Biztalk
> * IBM Websphere Business Integration
> 
> Note that I'm not pushing for or against the existence of this project.
> Just had some thoughts to toss into the discussion....
> 
> Cheers,
> 
> Simon
> 
> 
> On Thu, 2003-06-26 at 18:43, Jeremias Maerki wrote:
> > On 25.06.2003 19:59:19 Joe Germuska wrote:
> > > A CSV project seems a bit too narrowly focused.
> > 
> > I agree. IMHO the focus should be on any type of legacy, structured
> > ASCII files containing some notion of "record". It's very interesting to
> > have code around to easily convert such files to XML (SAX-Events), for
> > example.
> > 
> > > I guess I'd think it belongs in IO, of all the choices Henri 
> > > suggested.  In a sense you could associate it with Digester, but 
> > > that's probably not really the best place for it.
> > 
> > Having said the above, I think it doesn't belong neither to IO, nor to
> > Codec. I think it's a separate little project. The Digester idea is
> > interesting, but I agree.
> > 
> > 
> > Jeremias Maerki
> > 
> > 
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > For additional commands, e-mail: [EMAIL PROTECTED]
> > 
> > 
> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
> 


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to