Notes Re First Survey of Status
- START: Allowing online or email handin of reaction papers
- If you print off your reaction paper before noon on the day that
it is due and slide it under my door, that is a reasonable way to hand
it in if you can't make class. If you slide stuff under my door, you should
also email me to make sure I found it (if I don't respond within 24 hours or
shortly after class, whichever comes first, you should email me again).
- Online handins are a pain because marking stuff online is annoying.
So basically, anything I would get online would have to be printed.
In addition to the extra work of printing other people's stuff, some of
it wouldn't be formatted right for our printers and so wouldn't print
out properly. Also our printers sometimes run out of paper or have
problems with queues which make the whole thing annoying enough to
make one wonder about lowering someone's mark :-) However, it is also
worthwhile considering some of the available options and the problems they
have even before we get to the issue of printing
- ITS OWL:
- bad security. first rule of security is compartmentalization,
but ITS has everyone use one password for everything meaning that if anything
is compromised, everything is compromised. second security problem is that
machines are administered from off campus meaning if someone's home machine is
compromised, the whole system can be taken over -- in addition to compromising
confidential information, this could also cause machines to be unavailable.
problems impact both students and faculty. third security problem is how
is access to backups managed.
- lousy web interface.
- Email attachments designed to carry computer viruses. You have no way
of knowing if your computer is infected.
- Email unreliable. Some messages just seem to get lost, others get deleted
by mistake, others live a lonely life in some spam trap.
- Converting email to a printout is a slow and annoying process.
- I think ITS is using some sort of icky Microsoft package.
- Note: for obvious security reasons I don't use a web browser for email
(but also because the web based email systems are just so badly designed that
the thought of using one of them regularly makes one gringe).
- solves the password problem
- superficially appears secure (at least as much as any cloud app is that
is accessed by buggy web browsers), but unclear what their actual administration
policies are or where they find the money to get what is needed done.
- probably more reliable than email, but still no guarentee that it won't
go away next week or go behind a paywall.
- Store encrypted files in publically readable location on gaul where
they can be retrieved and decrypted elsewhere.
- From technical point of view, probably best solution, but it has significant `human' problems.
- People tend to choose bad encryption keys.
- While UNIX file security is a relatively straightforward system, people
who frequently use it eventually end up making mistakes and leaving something
they meant to protect exposed or don't make readable something they meant to
make readable. One problem is the system doesn't tell you if you made a mistake,
you have to continually test things yourself. Of course, this is a problem
with most computer-based solutions and a reason why I tend to advise people
to avoid computers where-ever possible -- programmers make plenty of mistakes
themselves, but expect their users to not make mistakes and to perfectly
understand the systems they are provided with.
- Anyway, worth considering the printout under door solution before
investigating a computer solution further.
- START: Post answers (or at least partial answers) to reaction papers and quizzes
- I post stuff from class that I put up on the screen in emacs. If I
forget to do so, just ask. If there is something that wasn't put in a file
during class that you think should have been, just ask (during class
where issue arose).
- The questions are all posted. The final will have different questions.
Study the assigned papers -- there is plenty to read there, so I have been
trying not to create a lot of side reading on top of the paper reading for
you all. Ask yourself:
- What is the point of this paper?
- How does the author demonstrate the claims of the paper?
- Are there useful examples provided? Can you think of any additional
places where the paper's ideas could be applied?
- Could you make use of the information in this paper?
- Does the author introduce any new terminology or way of thinking about
specifications, testing, or quality assurance?
- Am I comfortable that I understand everything the paper is saying?
Any confusing parts?
- How does this paper fit with other papers we have read so far? How does
all this stuff we have been reading fit together? If you are not already
familiar with them, concept maps ( wikipedia entry) are often recommended for organizing such
- Feel free to discuss the papers (but not upcoming reaction questions)
with other students to get further feedback on your understanding. A classic
comment is that in explaining something to someone else, one realizes what
one doesn't quite understand oneself.
- START: More guidance on learning objectives.
- The learning objective is for you all to learn more about issues
related to specifications, testing, and quality assurance. The hidden
agenda is 1) for you to also learn more about learning professional material
by reading articles that were published on it and 2) for the stuff you learn
to be practically useful to you. If you had had more time, it would have
been nice for most people to have undertaken more class projects, but given
the amount of project work in other classes, it is understandable that that
wasn't a reasonable option for everyone -- at least I hope you will find
things in this class that you can apply to make the other project work you
are doing better tested and of higher quality (even if it doesn't count for
a mark in this class, it is still practice of the ideas of this class and
should show up as a deeper understanding of the class material by the final).
The questions (reactive, quiz,
and final) marked in this class are mostly about are you reading and
understanding the articles (both individual articles and the collection of
articles assigned as a whole). Projects give you a chance to see what it
means to try and do testing and quality assurance better than you have in
- STOP: Dated articles.
- I don't expect we will see many more Fortran articles. Although it will
be interesting to see if people had more difficulty reading Fortran (a language
still in wide use) or statistics (another language still in wide use).
- But, in general, the early papers on an approach give the best explanations
and examples, because at that point the author is dealing with people who
haven't started to think about things in the `new way'.
While the technology keeps changing and the software libraries and applications
keep getting replaced, there is very little new in terms of actual ideas and
concepts in computer science. And what is new, built upon older ideas and
concepts and doesn't really make sense without them --
`Ontogeny recapitulates phylogeny' ( wikipedia entry) may be bad biology, but it is
still good education theory.
- START: Slides with key topics as study aids.
- Doesn't really fit objective above of getting you to learn to read
material and figure out what are the important parts. Those who are in
the 3rd or 4th year of university are in year 15 or 16 of a 16 year education
process -- if you can't read articles and learn from them now, what will you
be doing 5 years from now, when all the systems you have been taught have been
replaced with new stuff and there are new students who have been trained on
the new systems who will work for less than you have come to expect
(even though you haven't been trained on the new systems and want someone
to pay for that as well)?
- Note: if we are still doing testing and quality assurance
five years from now the way we are doing it now,
then computers will destroy civilization before global warming can.
- START: More open discussion of papers
- There is a limited amount of time in class which seems to always be
filled up, but if anyone has any questions or comments about anything
we have been reading or doing, just mention it and we will see how things
develop from there. Certainly I would like to see the class work as a group
to get deeper into the material, although so far I haven't seemed to trigger
that kind of participation.
- STOP: Going over so much detail on answers to reaction papers
- My intent is not so much to be talking about the answers themselves, but
to indicate how the answers should have been gotten to so that people can do
better in the future. But I would agree that we have probably spent too
much time on that by now (we are over 25% of the way through the semester
now), and will taper off starting in on that myself.
- START: Spending more time talking about things not covered by reaction
- I don't want to replace learning to read the paper with learning from
my lectures on the contents of the paper. But if other things are in the
papers that you are interested in discussing in class, please mention it
in class (see open discussions above).