How do geeks get it wrong?
I have spent a few days trying to use a new software package that has been developed for a large charity in Britain. The Norwich office had been told that the system would soon be going live,. So the idea was that some of us would be trained and then train the others. All very simple, except the system is very unstable.
How have the geeks who designed the system and wrote the software screwed up so badly. One possible reason is that it is based on a Microsoft product, this is not a good sign. The other issue is always that it is test by experts in a small localised environment, when it goes live it is being used by lots of people, probably more than the system can cope with. Surely anyone working as a computer scientist know about combinatorial explosion, when things grow exponentially rather than linearly. For example with 10 people using the system that is the possibility of about 1000 subsets of people do the same things make that 100 people and the things go batty, that 2^10, which is astronomical.
Another reason is that there are three main sets of people involved, the managers, the programmers and the users and often they don't talk to each other. This is made worse by the interference of salesmen. This goes roughly like this, the managers, who probably know nothing about computing, say we want to buy this. The salesmen say we can do that for you this quickly at this price in order to get the contract. Meanwhile the computer people know that this is not possible but who cares about them, the salesman want hsi bonus.
So the project gets delayed, meanwhile the managers eventually talk to the users who naturally want all sorts of modifications. The managers have no idea what is practical and what is by now not feasible. They got back and the company, who is now worried that are going to make a loss, say we can do this but it will cost a lot. Even if the organisation has its own geeks they will be ignored because they obviously know less than the consultants from the firm who are doing the software, cause they cost so much!
In the end the software is delivered late doesn't work too well AND if you area charity a lot of volunteers walk out because they are pissed off. Mean while the salesman has pocketed his bonus and with luck the other senior managers of the software company have looked after them selves.
Finally do not forget Microsoft who will pocket the licence money whatever happens. My simple advice is don't buy Microsoft and never trust a salesman and certainly never believe an outside consultant.
How have the geeks who designed the system and wrote the software screwed up so badly. One possible reason is that it is based on a Microsoft product, this is not a good sign. The other issue is always that it is test by experts in a small localised environment, when it goes live it is being used by lots of people, probably more than the system can cope with. Surely anyone working as a computer scientist know about combinatorial explosion, when things grow exponentially rather than linearly. For example with 10 people using the system that is the possibility of about 1000 subsets of people do the same things make that 100 people and the things go batty, that 2^10, which is astronomical.
Another reason is that there are three main sets of people involved, the managers, the programmers and the users and often they don't talk to each other. This is made worse by the interference of salesmen. This goes roughly like this, the managers, who probably know nothing about computing, say we want to buy this. The salesmen say we can do that for you this quickly at this price in order to get the contract. Meanwhile the computer people know that this is not possible but who cares about them, the salesman want hsi bonus.
So the project gets delayed, meanwhile the managers eventually talk to the users who naturally want all sorts of modifications. The managers have no idea what is practical and what is by now not feasible. They got back and the company, who is now worried that are going to make a loss, say we can do this but it will cost a lot. Even if the organisation has its own geeks they will be ignored because they obviously know less than the consultants from the firm who are doing the software, cause they cost so much!
In the end the software is delivered late doesn't work too well AND if you area charity a lot of volunteers walk out because they are pissed off. Mean while the salesman has pocketed his bonus and with luck the other senior managers of the software company have looked after them selves.
Finally do not forget Microsoft who will pocket the licence money whatever happens. My simple advice is don't buy Microsoft and never trust a salesman and certainly never believe an outside consultant.
Labels: Being riiped off again
posted by Unknown at
3:54 PM
0 Comments:
Post a Comment
<< Home