If we would do traffic like we do IT today every city would have their own graphic designer coming up with Stop and Yield signs.
To get from A to B, drivers would have to pass through $BigMac and wait 1 minute with the option to buy a 'meal' and continue on quicker.
I would argue that this is a side effect. The point is: code paid for with taxpayers money is not some proprietary bigco ip.
- less duplication of work (=spending of tax money) between structurally very similar communes
- less lock-in that keeps you bound to that one corp forever (=more incentives for them to deliver good work for the money)
- potential collaboration and improvements from outside actors (private persons, other communes, organizations, companies, ...)
The idea is basically: if we need to spend money for software anyway, why not develope it in a way that produces lasting value for society as a whole?
I agree with your general premise, though.
It's not only that others can grab your software for free, it's also the other way round - resulting in way fewer effort, cost, increased stability/quality etc for everyone.