Re: Irregular review procedures for IACR sponsored conferences

David Wagner <daw-usenet@xxxxxxxxxxxxxxxxxxxxxxxx> wrote:
Kristian Gjøsteen wrote:
Suppose I write a correct, well-written one-page note saying that Thm 2
and 3 of paper such and such are trivially not true as stated (giving
a trivial proof, of course), with the implication that any result
based on those theorems should be ignored. The note contains no other
content. Should such a note be accepted? Would accepting such a note
help us as a field learn from our mistakes?

I haven't read the original paper, so I don't know how central Thms 2
and 3 are to the paper.

I'll stress that I have only briefly looked at the paper. Also, I'm not
saying anything about what Markus has done, I'm only trying to construct
a hypothetical note about this paper that the program committee shouldn't
accept.

In the abstract, there is a claim that properties of something is
"mathematically provable". There are only three theorems. The first is
obviously correct, and says that a given map is a bijection on a set
of strings.

The second says that if you apply the map to an arbitrary but sufficiently
long string, substrings of the image are uniformly distributed. This
can obviously not be true, since the map is a bijection and there must
be a string that maps to the all-zeros string.

The third theorem says that the given map increases the period in the
image by some factor. Again, I expect this to be trivially false, you
could possibly exhibit some string s that maps to the all-zeros string
and show that s has larger period.

I don't see how the above three paragraphs contribute to anything.
In fact, I believe they only show that either the review was sloppy,
or the authors failed to take the reviewers comments into account when
preparing the final version. Probably, if you suitably qualify the
theorems with some statements about probabilities of such and such
happening, they will be correct. Markus can probably say more about that.

Just my take on things. I can completely understand if others would have
a different view. But in the long run, I personally feel like taking
this policy is better for the field. (When looking at any one case,
the policy might look pretty dubious. The benefits come not from any
one individual application of such a policy, but from consequences that
emerge when such a policy is applied consistently.)

I hadn't fully appreciated the benefits of such a policy, and I'm
grateful to you for explaining. But I am not sure that it should be
applied consistently, I think there could be exceptions.

--
Kristian Gjøsteen
.

Relevant Pages

  • Re: Oracle Connection Pooling
    ... > public String oracleFormatUrl(String prefix, String host, String port, ... > proprietaryException { ... > * Map from a resource attribute to a proprietary attribute. ... > protected void processCreateRequests ...
    (comp.lang.java.databases)
  • Re: The message must contain a wsa:To header
    ... When I setup the policy, ... at ApplicationMessagingWS.Dispatch(String messageType, String ... be used along with the Integrity assertion when the presence of the signed ... look for a certificate with this subject name in the certificate store ...
    (microsoft.public.dotnet.framework.webservices.enhancements)
  • Re: Mapped F Drive - group policy update problem
    ... again the drives maps correctly. ... users save work locally(not our policy) and on the network. ... If this setting is disabled or not configured (Windows 2000 Server Family ... set to map in the user properties of ...
    (microsoft.public.windows.server.active_directory)
  • Re: Regarding use of ArrayList as value in Hashtable
    ... say, for example,the value is a String array, with String array then we ... Lists, Integers, and Strings. ... An object splits the stream of characters into tokens of type: MAP, ... enum BContainers{ ...
    (comp.lang.java.programmer)
  • java.lang.NullPointerException..help me!
    ... public UploadedFile getMyFile() { ... // Populate the map from the input file. ... String line = null; ... // Create the array for the numbers which make up the rest of the line. ...
    (comp.lang.java.programmer)