change to flags and subtypes (was Re: rock-border cannot be presumed)

[Andrew Atkinson]

Wookey wrote

>>> >
>>> > OK - changing therion this way would be easy - but converting existing
>>> > data is difficult. I would be very happy if I could remove
>
>> subtypes, but
>
>>> >  it would be quite substantial change...
>
>>
>> It it really that difficult to write some perl conversion that
>> would usually
>> do the right thing?
>>
>> I think you are right that this was a design error and we should fix it if
>> at all possible. There are not that many datasets now - in the
>> future it will
>> only get harder...
>>
>> I think we should say that this will change in 0.4 or something. If we can
>> add the new functionality now/soon so that people can create new
>> data in the
>> new style, then that will make it easier to convert (because
>> there will be a
>> version that can cope with old and new data for people to use until they
>> have converted).
>>
>> I don't want to break my dataset any more than necessary, but if
>> changes are
>> necessary for long-term growth, and we can design a reasonable transition,
>> then it is probably worth doing.
>>

The difference and different ways to get the same thing confused me, I think
that although it would break my data, it would be better to change it now so
that the program can be developed coherently in the future.

Andrew

PS On a related topic, how does Therion now fit with survex, I have to say
that I am finding data set conversion a little tedious, and I have not tried
to for a big cave yet.