"Foundational" business software

>What I want to know is, when is the Free Software community going to get 
>serious about providing slick business/nonprofit "foundation software" 
>implementation?  

Possibly never.  As someone who developers in the groupware space (essentially, IMO, what you
are talking about), I am very very skeptical that Open Source can deliver these kinds of
solutions.  It requires tight integration between disparate projects... that is really hard
and almost diametrically opposed to how most projects view themselves.  Adhering to some
standard is an unequivicable loss of control of the direction of your project,  therefore most
developers aren't interested.  For all the pro-standard fervor in Open Source there is
precious little standardization in practice:  witness the CalDAV haters, the absence of
Kerberos support, etc...  [LDAP has become quite standardized] M$ has a real technological
advantage;  they can say "everything is going to use Kerberos and Active Directory" and
magically everything works together, the user doesn't have to enter their password to login,
to access a shared volume, to use intranet application A, intranet application B, and intranet
application C, etc...  The down fall of Open Source is that it is very good at producing silo
applications [albeit very good ones] and bad at seeing the big "enterprise" picture.

>This is the software that should seamlessly tie together 
>Free office suites with databases, mail servers, LDAP, etc., to perform all 
>the tasks that need to be done.

The problem is that "all the tasks that need to be done" is extremely open ended.  In my
experience (primary in the LDAP/Kerberos space and then in the groupware space) is that you
generally have two pots of users:  very simple users who think Google doc's, calendar, etc...
are the greatest thing every and meet *EVERYONES* needs.  And a higher tier of users who
*NEED* workflow, project management, and conflict resolution, etc... For the most part the
first group simply refuses to believe that the second group exists (see the dumbest BLOG post
ever - http://www.jwz.org/doc/groupware.html - for a *PERFECT* explanation of this attitude). 

> GNU Enterprise doesn't seem to be getting very far lately,

GNUe has been around for awhile, and I've never seen it used anywhere.

> and Compiere seems more like a cathedral than a bazaar.  

So what if it works?  Sometime you need a cathedral,  the bazaar is good at many things but
large complicated and *tightly integrated* apps are clearly not one of them.  You said right
in your letter to the editor: "I would not necessarily argue that said software *must* be Free
Software."  Don't move the fence post.

> Neither have a great deal of mindshare or excitement in the community.

Nope,  maybe because there is only a very small portion of the community that gets excited
about things like ERP and groupware?  Because most people fit into the first group mentioned
above and don't have those kinds of needs.

> This can't be emphasized enough.  This stuff is absolutely *critical* before 
> we can expect much more adoption on the business desktop.  

Yep.

>Until we produce  something that will tie together the free software stack, 
>these organizations have nowhere else to go but the commercial offerings, 
>which tie together the entire Microsoft stack, and lock them into it.

Yep;  M$+SQL + Exchange + Sharepoint do make a very compelling combination.  Then throw group
policies and PowerShell [Why on earth isn't there an Open Source answer to PowerShell at
least?!] and it is pretty intimidating.

Currently I'm working on PostgreSQL + OpenGroupware + MojoPortal and I'm pretty happy.  We've
built some really nice business front end applications in PHP -
http://docs.google.com/Doc?docid=ddv5htgd_14zrg6zm&hl=en

To get ubiquitous integration between CRM, Intranet applications, etc... you have to build on
a common core - and that is groupware.  Then you can access data with various clients,
applications, and devices and get a good cohesive user experience.

>I also believe that this is perhaps the *last* major roadblock to making 
>serious gains on the business desktop.  Technically, Linux works fine at 
>this point.  So do the databases, mail servers, and LDAP that would form the 
>foundation.  Linux and open standards have a lot of respect, but inevitably 
>organizations turn to what works and works without much fuss.

Yes, but you are thinking about it at too low a level;  users don't user databases, LDAP
servers, mail servers, etc... they use applications.  And making good applications requires a
solid and consistent core.  That core is a common layer on top of those services;  M$ does
this with Exchange and Sharepoint.

>Right now to build a business application on Linux, you have to go through a 
>lot of cryptically documented (if that) text on OpenOffice, KParts, 
>wxWidgets, or whatever you use.

I disagree;  there are good and solid application frameworks.  I've been developing in
Gtk#/Mono lately and the documentation is good, the product is stable, and the framework is
robust and complete.  Java can provide much of the same.