[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lmi] Terse list of valuable projects

From: Greg Chicares
Subject: Re: [lmi] Terse list of valuable projects
Date: Tue, 16 Mar 2010 20:10:28 +0000
User-agent: Thunderbird (Windows/20090812)

On 2010-03-16 17:40Z, Vaclav Slavik wrote:
> On Tue, 2010-03-16 at 14:45 +0000, Greg Chicares wrote:
>> advantages of xml, along with the disadvantage that its accompanying
>> schema is incompatible with free software due to licensing. It would
>> be better for us to use that xml format (without using the schema) 
> Just to be clear: do I understand it correctly that you want to use the
> XTbML format and not some home-grown format of our own? If that's the
> case, I'm unclear on the "without using the schema" part: does it mean
> not using its namespace? Or not including a copy? Or not consulting the
> spec/schema at all and reverse-engineering it from existing documents?

Probably all of the above, except that I think there's sufficient public
discussion that can be used because no one seems to claim copyright in it.
There are many documentary fields that are of no real interest to us.

Above all else, we want to avoid using anyone's "intellectual property".
That overrides every other concern.

Many tables are published by professional or regulatory groups, and it's
preferable for data-integrity reasons to use their authoritative published
data, e.g.:
AFAICS, those tables are in the public domain, yet in "XTbML" format.
Apparently there's an "XTbML" schema that's copyrighted, with a license
that's incompatible with free software, so we can't use that schema per
se; but there's a published reference here:
that I suppose we're permitted to refer to freely in order to write code
that reads these noncopyrighted tables. If you think that's questionable,
then I can just examine a table myself and write instructions telling you
how to interpret the parts we need.

I suppose it would be convenient to use "XTbML", if it suits our needs
well enough, and only if we can do so without violating the "overriding
concern" above. Our standard should be as high as the debian-licensing
mailing list's (which I believe is very high); if we don't believe we
could convince them that this is okay, then we should avoid "XTbML" and
continue to maintain our own 'actuarial_table*.?pp' with the binary
"table manager" files from soa.org that are free of "XTbML"--but convert
them into an xml format of our own (for cross-platform portability, and
also human readability, as well as ease of maintenance).

If you feel that a different xml layout would be superior, that's okay
(and I wouldn't be at all surprised if you do). We could download stuff
from soa.org (in whichever format we find more convenient, subject to
software freedom) and convert it to a better xml format. (I tend to think
xml is the best choice, but tell me if you think otherwise.)

Some tables are proprietary to a particular insurance company and would
be distributed along with lmi to its end users; these generally aren't
given to us in any particular format, so we can put them into whatever
xml format we like. Probably we should use one format for all the tables
that lmi routinely reads.

Hmmm...I just looked at an "XTbML" table, and...the more I think about
this, the more I suspect your vector format for our '.db4' and '.tir'
files is better than what they've done....

reply via email to

[Prev in Thread] Current Thread [Next in Thread]