incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mattmann, Chris A (388J)" <>
Subject Re: [VOTE] Apache cTAKES 3.0.0-incubating RC5 release
Date Fri, 25 Jan 2013 18:59:07 GMT
Hi Guys,

On 1/25/13 10:29 AM, "Benson Margulies" <> wrote:
>> Sorry, that's really what I meant: to think about that file as to
>>whether it
>> can do any harm and how to determine its safety.  I haven't looked at
>> file, but it sounds like you know it can do harm.
>Depends on what you mean by 'harm'. A model could exhibit hostile
>behavior, intentionally giving wrong or biased answers for certain
>inputs. Some people would call that 'harm'. Others would focus on
>malware, in which case having the model be a big file of numbers that
>gets turned into a big array would indeed be defined as harmless.

So, this is all conjecture. I don't think either of us are qualified
Benson to comment or even remotely suggest the harmful nature, validity,
etc., of a medical model, or set of them developed by institutions who
have accepted the risk (well beyond the ASF's willing accepted risk, etc.,
for *software*) of leveraging the model for scientific study and
prediction. Those institutions go as high up as the National Institutes of
Health, the Department of Health and Human Services, etc., etc.

Here's what it boils down to -- I get Roy's point, but does everyone else
get Roy's point? Roy isn't going to *sanction* wearing big ASF hat the
inclusion of things that the ASF isn't willing to accept risk on. The good
news? The ASF is NOT a top down organization it's a bottom up one with
minimal centralized oversight. That being said there's a subtle hint Roy's
text on record that describes what *individuals* as part of a PMC (who are
the ones that *release* the software) are, and are not able to do, and
what risk they are and are not able to (implied) take on, and perform,
within their community. This isn't the ASF coming top down with a policy
(good luck every getting one) -- but it's something that has happened in
the past with respect to binary "convenience" bits that even people like
Roy say can go on the server.

So, why don't we let this sit for a bit, and cTAKES community (and
mentors, please speak up) if we are still wanting to proceed down this
path, then let's do so, and I think we're going to be OK.

See post from Jukka Zitting and I on cTAKES lists:


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message