With Rupesh's invitation to subreview a HiPC paper and his tips (below) started my 1st offical review.

Think on these lines:

• Critic: Some technical points you could be suspicious / careful about. Point those out, mention your reason for suspicion (without calling it suspicion) and ask.
• Audience: Some points you are curious about. Ask those.
• Advisor: Some points you can suggest to be done better.
• Collaborator: And then some points may come from a high level thinking.

Rajesh finished the review in 4-5 days. Feels:

It was a different experience I thoroughly enjoyed it. But a time-consuming process. .. I had to night out to type in all or full parts -- Ideally, I should have typed in as read.

Rupesh points

.. as we read in document viewer highlight.. writing comments on highlighted portions as we review. .. then go through it again to type out my review, which allows me to reorder the comments based on importance and also to connect some comments which are related (e.g., a certain advantage mentioned in theory is not brought out in experiments). ..The review becomes more important when we are negative. It is crucial to not only highlight issues, but also to suggest ways to improve the work / paper (constructive criticism), so that the authors have a better chance of acceptance in their next submission. The authors then appreciate the review process / conference better, and also reciprocate when they get to review.