November 14, 2005
first draft of proof paper due Monday 11/21
Instead of Thursday. Also, don't forget to add your numbered definitions, theorem statement and figure.
Posted by jones at 06:08 PM | Comments (0)
November 11, 2005
You might consider using Latex
Latex is a very old document preparation system that has its problems. However, latex is still the easiest way to prepare a manuscript that includes a fair amount of mathematical notation. Latex is available for windows (as miktex) and *nix as latex.
There are many online references for using latex, including a list of symbols and a guide to equations in latex.
Latex documents start of as text files. I've created a generic paper text file (and the resulting pdf) that you mind find useful as a starting place.
Posted by jones at 03:34 PM | Comments (0)
November 10, 2005
Definitions papers, in general, need work
The definitions papers need some work. Very few of the papers included formal definitions and even fewer included a formal statement of what was to be proven.
For the first complete draft of the paper, I want to see: at least 3 numbered definitions, at least 1 figure explains somehting in your paper and your theorem as a numbered theorem.
Posted by jones at 05:51 PM | Comments (0)
October 10, 2005
Ideas for the proof project
Here's a few ideas for the proof project. They come from two areas: the class and my research area. Both of these topics are not original new proof problems. That's ok. If you can't think of a proof project that is related to your research then it is going to be hard to dig into a proof in 2 months and do something original. Just don't look it up on Google and rewrite a proof you find there.
Formal methods and model checking:
- What properties of omega-regular languages are decidable? how does that compare to regular langauges? Can you prove something like Rice's theorem for functions that classify based on omega-regular langauges?
- Prove that the use of an "empty buffer" to get a finite set of partial concretizations of an abstraction state generalizes to "n special buffer types" that can be replicated and still retain a finite set of partial concretizations. Also derrive a bound on the number of such concretizations. This is based on a paper I just finished. This would be hard but interesting and I would be interested in advising an MS student working on this topic. You would be expected to get a good start in CS 611 but may not finish (by the way, everyone gets that lattitude on their own proof projects. I value "interesting" more than "safe" and the grading will reflect that).
Proofs from theoretical computer science:
- Homework 3.14 and the generalizations on the arithmetical hierarchy (There are 3 on page 65).
- Homework 3.32
- Homework 6.22
- Homework 6.24 (see also 6.23)
Posted by jones at 10:14 PM | Comments (0)
September 27, 2005
What you all had to say about what makes a good rigorous paper.
I have finished grading the reports about "what makes a good paper" Grading was pretty lightweight. This is a tough thing to grade so I graded on evidence of concrete thinking and whether or not your opinions about "what makes a good paper" seemed related to the paper you read for class. Grades should appear in the gradebook soon.
I thought you might be interested in the consensus for the class:
What makes a good formal paper...
- know your audience
- use examples as appropriate
- don't use complex notation to present simple ideas
- motivate the topic so the reader will know why they are suffering through your paper.
- give some clues about why your proofs work but keep your proofs tight and terse
- Figures are good, unless they show something obvious in which case they are useless.
- Not everyone cared about style, grammar etc, but some cared a lot.
- Use a typesetting program, such as latex, to format your paper nicely.
Posted by jones at 05:48 PM | Comments (0)
December 03, 2004
Grade breakdown for the proof project.
As you know, its 20% of your overall grade for 611. The allocation of points to parts of the project is given in the body...
4% ~ 2 reviews (@ 2% each) 9% ~ 3 preliminary drafts (@ 3% each) 7% ~ 1 final draft (@ 7%) 20%.
Posted by jones at 10:22 PM | Comments (0)
Peer-review of proofs
You should turn in your proof review via email, the web site or the homework box or a combination. If you turn in your review in the homework box, put your name on a removable cover sheet and the title of hte paper you are reviewing on the review.
Turn it in by 10am and I will bring them to class for distribution.
Posted by jones at 08:22 PM | Comments (0)
November 29, 2004
Proof turn in.
[Syllabus for 611] You should turn in your complete draft of your paper, including the definitions and proof by Tuesday night at midnight. You should turn in your paper, as before, using the class conference web site. Use the same author-account you used before. You will get your papers to review on Tuesday night or wednesday morning.
No homework for Wednesday.
Posted by jones at 08:40 PM | Comments (0)
November 18, 2004
First version of Proof now due 30 November
I am going to delay the first version of the proof project to Tuesday the 30 th of November. Tuesday will be a hard deadline because I will mail out the papers for review first thing Wednesday morning. This will free up some time over the break for you to, well, take a break while ensuring that the proof first drafts are good quality.
Posted by jones at 03:30 PM | Comments (0)
November 15, 2004
Reviews are available.
[read your reviews] I've sent acceptance mail to everyone. Come get your reviews. Note that you have two accounts on the conference page: one as a reviewer and one as an author. You'll need to use your author account (username and password in your email and picked by you) for this one. There are a couple people missing a review for various technical problems. I will get those mailed out tout de suite.
Posted by jones at 11:41 PM | Comments (0)
November 05, 2004
Where to submit your definitions papers
[Refereed Abstract and Paper Submission] visit the above link and submit a paper. A few notes.... Don't foget to put in an abstract, it can be short ... The submission process is not very intuitive ... you have to create an account, submit your abstract then re-login to submit the paper ... submit the paper using "document submission" link at the top of the "edit an abstract or previously submitted paper" page ... Papers must be submitted as pdf's ... See pdf995 on windows and dvipdf or dvipdfm for latex users. It'll be fun. ... you will get a confirmation email.
Posted by jones at 09:02 PM | Comments (0)
November 03, 2004
Thoughts about the paper for Friday
I've had a few questions about the proof project. Here's the answers.
If you think your definitions and theorem are going to run about 3-4 pages, and maybe 2 pages, then you are on the right track.
Take the risk of maybe not completing the proof rather than taking the risk of proving something that might be useless but easy to prove. Grading will reflect my interest in maybe not finished but interesting difficult problems over easy, finished and unimportant problems. Important means important to your research goals.
Try to think of something that's related to your research goals and interests.
Posted by jones at 08:52 PM | Comments (0)
November 02, 2004
First part of proof paper due friday
Look in the syllabus (see "course documents" to the right) for details. Basically, you need to come up with the theorem to be proven and the definitions needed to understand that theorem.
Posted by jones at 04:30 PM | Comments (0)
December 02, 2003
Final Part of Proof Project
For the final part of the proof project you should turn in the following: The final version of your proof paper, your original scratch work, your reviewers' scratch work, your responses to the reviewers comments, an explanaition of what you did to address any major differences between your scratch work and your reviewers' scratch work.
Due the last day of classes and no later.
Posted by jones at 10:15 AM | Comments (0)
November 24, 2003
The Death of WIMPE?
WIMPE appears to be dead. If I ressurect it, I will post an entry. Otherwise, email your reviews directly to me.
Posted by jones at 04:32 PM | Comments (0)
November 17, 2003
Papers are ready for review
You should have recieved one or two copies of the following message:
I have assigned papers to reviewers. You can find your papers in your PC member directory. This is located at a url sent to you after you registered as a PC member for the class. There's a username (an integer less than 25) and a password that you set. You'll find links to your paper and your review forms. Don't forget that you need to include a reverse engineering of the proof into scratch work as part of your review.
If you didn't, then wimpe doesn't know your address and we have a problem. Send me mail.
Posted by jones at 03:39 PM | Comments (0)
November 06, 2003
Register for Proof Project Reviews
We are using WIMPE to manage the review process for the proof project. Use the CS 611 Proof Project Program Committee Registration Form to register as a reviewer. Eveyrone in the class must register by November 14th. We are ignoring the keywords, so you can ignore them too. You'll upload pdf versions of your papers using the author registration form. (Instructions for creating pdfs and uploading your paper).
Posted by jones at 10:22 AM | Comments (0)
Submitting Final Proof Papers
Papers for the proof project are to be submitted using WIMPE. You will be required to upload your paper in the Adobe Portable Document Format (PDF). If you are using latex, you may use dvipdf to generate a pdf document. If you are using Word, you may download and install PDF 995 to print pdf files.
The registraction form requires quite a bit of contact information. We only require your name and email address, you may put bogus information in the other fields. We aren't using keywords, and you are not required to check any keywords.
When you get the confirmation email, be sure to save the URL for future use.
The deadline for ontime submission is Friday November 14 by midnight (which means Monday by 8am).
Posted by jones at 10:12 AM | Comments (0)
October 13, 2003
Review form for Definitions
The review form for the definitions project is now available at the above link, or in the body of this entry below.
Be honest but curteous in your reviews. Its my experience that grad students write the meanest rudest reviews. Don't do that. Your review will be graded on its depth of thought and level of politeness. Your definitions paper won't be graded based on someone else's review.
Review Form for Part 2 of Proof Project
---------------------------------------
Reviewer #
Paper title:
Paper author:
What is the motivation for this problem?
Comment on the clarity and precision of the definitions. (Could you
tell what was being defined? Did you understand the definition? Did
the definitions include enough intuition and examples to be
understandable? Were the definitions precise enough to remove
ambiguity?)
Did you understand the theorem to be proven?
Constructive comments about style, organization and anything else for
the authors:
Scores:
5 = agree
3 = sort of agree
1 = disagree
Motivation was clear 5 4 3 2 1
Definitions were readable 5 4 3 2 1
Definitions were precise 5 4 3 2 1
All terms in theorem defined 5 4 3 2 1
Posted by jones at 10:46 AM | Comments (0)