[qrcodetag size="150" link="false"]We managed to get our first scientific paper accepted at the PESOS workshop at ICSE 2011 on Hawaii – which is obviously a very nice location for Surfing and other leisure time activities. However, we are there for the workshop . Our paper discusses Tweetflows and their application for workflows. Those of you who are interested in Tweetflows, can take a look at the submitted version of the paper. We will prepare the camera ready version in the next weeks.

We got the following comments on our work:

—————————- REVIEW 1 ————————–
PAPER: 3
TITLE: Tweetflows – Flexible Workflows with Twitter

The paper describes a collaborative platform based on Twitter and SOA. Basically, authors
describe their experience with the implementation of human/machine hybrid service compositions blending human skills (used for translation) and an IT-infrastructure to manage translation requests embedded in Twitter.
The idea fits well with the workshop topics and the audience will benefit from the authors’
contribution and the ensuing discussions.

PRO:

The idea is good and sounds interesting. The case study proposed is effectively a good example where human intervention is needed to achieve high-quality results on top of a computer-driven system. The usage of a recent, widely used tool like Twitter also extends the interest and fruition of the experiences reported by authors.

The paper is globally well written: first a verbose, human-readable description of the idea is given, and then described into more technical details about the implementation. Related work is ok.

CONS:

Consider adding an image with a schematics overview of the infrastructure used.
There are several typos, stylistic standards, and naming conventions that need to be fixed. For example:

- introduction: first occurrence of “SOA” should be written in full, and then robustly
referred as “SOA” (there’s still a “Service Oriented Architecture” at section 4.1.3
- introduction: avoid to begin the last sentence with “And finally”. “Finally, ” is enough.
- section 1: “Such a lightweight” exceeds the column size
- section 2: several typos (“must must”, repeated; “ikangai solutions uses”, remove one s; etc.)
- check that the hyphenation used is correct
- find a better solution to the repeated “(human-provided)”
- “adhoc” should be “ad hoc” (even better if in emph{“ad hoc”}, since it’s Latin)

—————————- REVIEW 2 ————————–
PAPER: 3
TITLE: Tweetflows – Flexible Workflows with Twitter

The paper is presenting an innovative approach which uses the Twitter micro-blogging service as the technical and social platform for service-enabled workflow management. The proposed solution reuses the infrastructure provided by Twitter in order to define simple workflows, bind each activity to human or software services and execute the workflow. Twitter primitives are used in order to publish, bind, invoke and request the status of a service and the connection with SOA-principles is illustrated sufficiently. A case study supporting the proposed approach is presented, drawing from the experience of the authors. The proposed approach is promising with respect to how coordination between human and software services can be achieved and enacted for lightweight workflows.

Despite the fact that the work is of good quality, the following issues need to be addressed:

1. Despite the use of the term ‘workflow’, the presented approach supports only sequential activities. How more structured/complicated activities can be handled from this approach should also be addressed, even partially.

2. Related work should also cover approaches like BPEL4People, that also aim for collaborative human and software service process definition and execution.

3. The paper needs to be proof-read again. The @ symbol in page 2 for example is not rendered correctly, the word ‘later’ should be replaced with ‘latter’ in p2c2par2 etc.

Furthermore, and given the way that the proposed primitives are structured and access service resources, the authors could also consider RESTful services as a more appropriate setting for extending their work, see for example:
C. Pautasso, O. Zimmermann, and F. Leymann, “RESTful web services vs. “big”‘ web services: making the right architectural decision,” in Proceeding of the 17th international conference on World Wide Web, pp. 805-814, 2008.

—————————- REVIEW 3 ————————–
PAPER: 3
TITLE: Tweetflows – Flexible Workflows with Twitter

The paper presents an interesting work, where the use of Twitter as a comminication and
collaboration platform for the realization of workflows over mixed human-based and software
services. Specifically, it provide te necessary primitives to publish services, discover services,
and even invoke services through Twitter. The relation to “traditional” SOA is discussed and some
results of experimenting with an application are presented.

The paper is well written and presented. The use of Twitter (or, however, similar in nature
platform) for the advertising/dicovery of social-based services seems an interesting idea and the
authors also demonstrate interesting consequences of that. Specifically, this regards service
discovery, the use of external REST for inputs/outputs. In my opinion, the use of Twitter also for
the service execution, monitoring, and other worklfow-related activities, is bit naive and requires
further elaboration. The Twitter approach is too public in these regards; the important information
about who,how, and what executes will be difficult to hide, but in some domains this may be not
very critical.

In overall, I believe this paper may foster an interesting discussion regarding the methodology of
realization of applications actively involving human-based services or human activities.

your happy ikangai science team

Tags: ICSE 2011, PESOS, Tweetflows