IMC will use double blind submissions, per the following guidelines:
Author names and affiliations must not appear on any submission.
Identifying information such as grant numbers must not be included on submissions.
The text of the submission must refer to the authors’ own previous work in the third person.
IMC recognizes that measurement research sometimes leverages unique capabilities that are publicly known, but not available for general use (e.g., measurement platforms, unique vantage points, tools, datasets, etc.). Revealing access to these capabilities or knowledge of them which could be interpreted as only being possible for their originators/authors, will de-facto de-anonymize a submission. However, attempts to anonymize these capabilities will not allow authors to leverage previous work on developing sound measurement capabilities, leaves reviewers without a full understanding of the new work, and may also be essentially impossible in some cases. IMC recognizes an exception for authors referring to such capabilities—even though this may de-anonymize submissions. The PC will not penalize submissions for these breaches of anonymization. We make several important comments on this exception:
While this exception will be important in some cases, most submissions can be effectively anonymized and therefore invoking this exception is expected to be relatively rare.
This exception only applies to text that discusses capabilities that cannot be blinded without significantly degrading the submission. The submission is expected to otherwise be anonymized per the general guidelines. Frivolous use of this exception to de-anonymize submissions due to laziness or in an attempt to curry favor may be held against a submission.
The expectation is that this exception will mostly be invoked for existing lines of research whereby authors are leveraging some well-known, but non-public facility. In most cases, first time use of a tool, vantage point, dataset, etc. will not be known to the community and, hence, can be reasonably well anonymized.
Even though a specific vantage point is unique—e.g., monitoring the border of a specific University—that specificity is often not important to the submission and using a generic label like “a medium size European University” will suffice.
A unique vantage point—e.g., a large network telescope—need not be named, per se. However, authors should not elide important details about the vantage point in the name of strict anonymization. For instance, while the organization running a network telescope need not be given, authors are free to convey the size and scope of the telescope—even if these details give a strong indication of the authors’ organization.
Consider a measurement app for measuring mobile phones whereby neither the source code nor the collected data have been publicly released. A first submission on the app and data can be readily anonymized. Subsequent IMC submissions that leverage data from the app or which extend the app may name the app in order to leverage the previous work—even if these details give a strong indication of author identity.
A public dataset—e.g., RouteViews—can and should be named. Since the dataset is public anyone can utilize the data and therefore this has no bearing on the anonymity of the submission.
A public or a widely accessible measurement platform—e.g., RIPE Atlas—can and should be named. Since the platform is widely available, naming the platform has no bearing on the anonymity of the submission.
While the program committee will allow for breaches of anonymization to name unique capabilities where important, authors are encouraged to be judicious in their use of this exception. Specific questions should be forwarded to the program committee chairs.
The IMC Steering Committee has also released a statement on double-blind review that can be viewed on the SIGCOMM website.