Paul Vecchiato reports that the City of Cape Town and the State IT Agency (SITA) plan to jointly develop an open source solution ... to manage their libraries. They estimate the development costs of the project at between R3m and R4m.
The language of the quotes seemed to me to indicate that they were planning to develop (as in, write) the software themselves.
That frightens me, as do most announcements that people are going to go it alone with open source. Why? Because I don't want to see them fail and have the blame shifted to open source in general, as opposed to poor planning, lack of skills, impossible expectations, and so forth. Oh, it's possible to fail despite not having any of the above, but those hits, being valid, are necessary.
I emailed Paul questioning the development strategy, and he responded that they are still in the early days of deciding what gets done by whom.
I took a look at the feature-set of PALS (the library system they're replacing), and it doesn't look too scary to me. The potential for improvements and new features is huge - they might even make going to the library something I'd ever consider.
Development costs being nebulous, this can mean that the cost of the actual software development (and related work, such as end user documentation, migration documentation, and so forth) is generous, or that it may not be particularly enticing to development companies with open source experience, as large chunks of the development cost may include the costs of transporting all librarians in the Western Cape to training on the new system multiple times and so forth.
In any case, I really hope that this goes to tender rather than being developed in-house. And similarly, I hope that whoever's tender gets accepted actually understands open source and how this makes this project different from just another bespoke development project.
And I hope everyone knows that the key to being open source is actually distributing the source code under an OSI-approved licence (BSD and LGPL seem appropriate to government work) to all comers. In other words - not another Cape Gateway which is/was only source-available and only to some people.
We have a (slightly smaller) issue in our department. The IS types are going ahead and doing all sorts of analysis and so on to decide what features are needed for a new asset tracking and management tool. Now, the first thing that struck me is, "this seems like something other people have _probably_ thought about", and after a cursory glance at sourceforge it was evident that there already was a feature rich, actively developed, mature looking project to accomplish a lot of the things i am guessing they need (i mean, asset tracking is pretty boiler-plate organisational stuff). Perhaps the interface is not exactly what they need, and maybe the feature-set is smaller or larger than they spec out, but it seems a pretty good starting block. Jeff Norris from NASA has some interesting things to say about using existing frameworks to build up from at http://phk.freebsd.dk/pubs/ieee.software.pdf My point is, this is another thing that seems to be something other people (probably OSS types) have looked at, and have started to approach. The whole point of FLOSS is shared base of ideas. Perhaps even if there is a limited amount of work available centering on this kind of task, it is still 100k of labour in the right direction (thinking with government money). For 3-4m you could hire (at max) 33 developers for a year @ 10 000 ZAR per month. Perhaps not the best salary in the world, but if they are free-lancing then it is a good freelance contract. Granted the cost does not only come in development, and you would need some kind of management structure, but this kind of problem strikes me as a 5-7 developer 6-8 month initial development project, and then support and extension. And, if you start with an existing project, or identified building blocks, you have already knocked 3-4 months off the top. My point is, I hope they are not being silly about the development process, and looking to get good bang for their buck, while, in the OSS philosophy, joining, or creating a community of people with common interests. I think you can create /supplement an active, productive community quite nicely with 3-4m ZAR.
To carry on ranting...This whole develop from scratch is not a good idea, unless it is very clear that there is _nothing_ suitible out there...And...imho the develop from scratch mentality has grown out of a time in the software world when we were dominated by vendors with their custom solutions to everything, and not in a world where people get together and build on top of other people's solutions...
http://www.oss4lib.org/
One would hope they have a good reason to not use Koha (http://koha.org).
Just for clarity. The City of Cape Town is not looking at developing from scratch. We have done extensive research on the various Library Management Systems initiatives around the world. Koha, while very popular, is unfortunately on too small a scale for what we are looking at. The Evergreen project, which is and OSS development of a LMS for PINES (Public Information Network for Electronic Services) in Atlanta seems to be more appropriate for our needs. They are designing for double their current requirements of 249 libraries and book mobiles over 123 counties in Georgia. They have 1.3 million registered users and 7.7 million catalogued items, circulation figure of 15 million. This is a far closer match for our requirements.