[SystemSafety] Bursting the formal methods bubble

[Mario Gleirscher]

Dear Les, and all,

what a true and insightful story, loved "if you aspire to
professionalism, be honest, accept criticism and harden up – unless of
course your being reviewed by bozos."

As a researcher, high quality peer review is decisive though vastly
underrated and scarcely incentivized. So sad!

But let me briefly comment on SW peer reviews:

No one has to convice me of the advantage of a proper SW peer review, be
it selective walkthroughs or comprehensive systematic inspections.

Anyway, Les' blog post reminds me of a nice programming team I've been
working in when jobbing as a freelancer quite a few years back. The
domain was basically non-critical, however, for imho good reasons, the
chef architect decided to introduce some kind of light-weight code peer
review to be supported by a small web app integrated with trak (an
SVN-aware SW dev/mgt wiki, not a bad tool actually).

Consequently, I became thrilled to get my code, which I had to "inject"
at some quite critical points into the system back-end, finally read by
someone else from more experienced team members, with the prospect of
learning something from their briefest comments and, of course,
improving the system and not having to wait for non-existing or hard to
craft and conduct tests (a fair win-win so to say).

However, nothing like that eventually happened because most of the
programmers seemed to quite quickly develop diplomatic reluctance
against this potential little helper for, I think, reasons such as "I
don't have time for this, I don't want to critize my colleagues, I don't
want to gain bad reputation, this is not organized or backed enough by
mgt., ..." and many more I suspect, ... all reasonable but sad for me
and the team! I have to clarify, there was no money and no obligation at
all in this domain to acquire external reviewers. I also have to say
that I have never asked the chef architect why he has not been insisting
more strongly on establishing a imho pragmatic cultural change in the
development team.

In the meantime, I've learned from other cases and been reported from
colleagues that even start ups in non-critical domains manage to
establish the described culture. It is also clear and very important to
say, that tooling (FindBugs, lint, ...) and IDEs (JetBrains et al.) have
gotten much much better since Fagan times where people had to perform
bugfinding from their checklists manually(!) However, for critical
domains, I hope that we can someday apply a similar social mechanism to
introduce light-weight, more rigorous correct by design methodology
(maybe with FMs under the hood) into software development culture
helping developers constructively critizing their work and improving it
instead of getting annoyed with tedious tasks and using non-compliant
tools.

My long 2p as well :)

Have a nice day,
Mario



On 08.11.2017 10:38, Les Chambers wrote:
> http://www.systemsengineeringblog.com/extreme-review-a-tale-of-
> nakedness-alsatians-and-fagan-inspection/

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5053 bytes
Desc: S/MIME Cryptographic Signature
URL: <https://lists.techfak.uni-bielefeld.de/mailman/private/systemsafety/attachments/20171109/5edba79c/attachment.bin>