Q&A: In some computing projects do some people introduce far more errors than they ever fix ?
Question by michel foucault: In some computing projects do some people introduce far more errors than they ever fix ?
Was “ordinary language” well-suited, or ill-suited, to Collins project ?
•by How stupid am I
There would need to be “protocols”, call them a language or languages if you like. There would need to be protocols that ensure success of the project.
•by secret_s…
All you can do is try and record what happens and look at the record later and try to extrapolate to the real problem. I don’t think the language matters. 75% mutual intelligibility might be adequate.
•by Shariput…
•Are we talking about communication problems ?
•
•by sherlock…
We’re talking about protocols that are fit for purpose. If the protocols aren’t fit for purpose, extra effort will be needed from us, if we’re still trying to deliver a system that is fit for purpose.
•by Socrates…
You can only try and log any progress or lack of progress that you make, so you or someone else can come along later and extrapolate to the real problem.
•by remyr
So, hypothetically, what sort of a log are we talking about, supposing one was kept ?
•by The Path I Choose
I think ordinary language is peculiarly ill-suited.
For example, just saying “Collins project” – person A will immediately want to jettison all legacy systems, person B will not want to differentiate between Collins systems and legacy systems, and will see Collins systems as interoperable with legacy systems and legacy systems interoperable with Collins systems.
•by Waikiki Kahuna
“Ordinary language” will work, if people keep to the original plans.
•by whistleb…
And keep to the already existing model-based testing test models and don’t produce forged versions of these.
•by witnessp…
And generally speaking keep to the spirit of the formal and perhaps informal documents that were produced even before contract signing. Wouldn’t some of the writers of such documents have already tried to answer these questions, for us as it were ?
Example- they might allude to a suite of CAI (computer aided instruction) computer programs that they know about and helped prepare.
Example- they might allude to Collins models already being prepared at the time they wrote.
Example- they would be aware of the desirability of interoperability between artefacts produced pre-contract signing and artefacts produced post-contract signing.
•by workplace psychopath
Whether using “ordinary language” or not, I think people might have to “keep up with the recent literature”.
The recent literature would alert us that there are not enough quality inspectors in the Universe to ensure a good product if the people producing it don’t want to produce a good product or don’t know how to.
•by counters…
Whether people are using “ordinary language” or not, some of them are likely to be introducing far more errors than they fix.
The project then look like the top floor of a Manly ferry, with people boarding, while stragglers are still getting off the ferry. It works well for Manly ferries. But on such a project as this there is the spectre of
the few residual bugs in the system being removed (the few stragglers still getting off the ferry), while the large number of people getiing onto the ferry represents the large number of bugs that are being introduced to the system (because if people come up with changes, the changes may be ill-considered and many of them probably are).
•by Easter Bunny 72
I don’t understand how this could happen (error being introduced – in large numbers) when people would have already published plans specifically to stop this occurring.
•by 18 ’til I die
So whatever “protocols” or “ordinary language” we’re thinking of, it won’t be much use unless it ensures that the good plans, already agreed, are not unilaterally changed.
• by michel foucault
I don’t care whether it was unilaterally or not. Are you suggesting agreed plans were then changed ?
•by Googy Egg
I know we’re only guessing that might have happened, but if agreed plans were changed, the project would be “all over red rover”. That would be like changing Patriot missile software, or changing software for the Mt Erebus flight, or changing software for a manned space mission.
•by Jojo207
So we need a “protocol”, an “ordinary language”, that takes the guesswork out of it.
Also I’m kind of confused becuase wouldn’t there be massive overkill in terms of the amount of cross-checking
able to be performed ? So people can guess all they like, I don’t see how they can botch the system just by leaving out a few cross-checks.
Best answer:
Answer by Houston, we have a problem
– Ah yes, a QA Directors nightmare.
Give your answer to this question below!