Re: [Hampshire] basic job costing / call charging / support …

Top Page

Reply to this message
Author: Daniel Pope
Date:  
To: hampshire
Subject: Re: [Hampshire] basic job costing / call charging / support logging system needed / project costing
On Wed, Apr 09, 2008 at 11:00:40AM +0100, Steve Kemp wrote:
> > It's not open-source though, not yet and possibly not ever. It
> > represents a sizable amount of development and I'm not going to open
> > source it before considering what return I can get on that investment.
>
> I'd expect that if it were good *and* it were web-based it would be
> relatively straightforward to sell - either hosted, or per-server
> licensed.


Yes, it would, but it would require substantial further investment of my
time to generalise it for the diverse needs of the market and then to
promote it.

For example, I have a client who wants to use the invoicing component of
it. To meet their needs it must divide payments into an initial lump
sum and 11 monthly payments. My payment terms are a hard-coded 30 days.

The formatting of the PDF invoices generated by the system is
hard-coded; for an off the shelf product this would probably require a
choice of layouts with customisable logos, headers, footers and so on.

Open-sourcing it could help it grow into a more useful product, or it
could take away its uniqueness and reduce its usefulness to me as a
showcase and as something that clients would pay for me to adapt for
them.

I don't hate freedom or anything, it's just for freedom to make money
requires rather shrewd marketing.

> With a desktop component too though? That sounds like it would only
> work on Linux/Unix desktops, and I'm curious what kind of market size
> that would have. (I could accept that this is just optional..)


It does only work on Linux/Unix desktops. The choice of KDE was because
that is what I'm most familiar with and it met my needs. The application
itself is rather simple - 600 lines of Python and two Qt dialogs - and I
could port it to a variety of other platforms, such as Java, .NET, PyGTK,
AIR or even just AJAX if I wanted to commercialise it.

The only complicated bit is idleness detection which is apparently rather
hard and very platform-specific, but immensely valuable. My current
approach is equivalent to

watch -d grep i8042 /proc/interrupts

which works very reliably, but with PS/2 mice and keyboards only. There
are X extensions but that's a whole world of pain to implement in a
PyKDE app.

Dan