Good things - phploc examples and going over user's own code
Bad things - a bit rambling and not enough focus on "how" to review versus doing the actual reviews
Showing how to review code is generally as effective as showing really bad PHP code, I'd suggest limiting it to just one (really bad) example of crappy open source code and then have a checklist for "how to review" and "bad/good practices"
It as great comedy, and fascinating, but I also wish there had been more organization and step-by-step. That said, the three speakers clearly show the tension and compromises that you have make between certain goals (security vs architecture).
The large size of the auditorium as well as the bad quality of the audio system really made this session hard to follow.
Also, it is understandable that the quality of the talk will only be as good as the provided examples, and some of the samples weren't that great (who wants to review good code??)
Though comedic at time, the chemistry between the speakers wasn't great (maybe due to the noise in the room, or problems with the microphones). I think it would have been more effective with just Stefan doing the review, and possibly Sebastian talking about the code metrics.
The first part of the talk was useful (open source projects), but the second part lost my attention.
Terribly disappointing. The sound problems were awful (I could hardly hear anything the speakers were saying) and it didn't help that the trio seemed to be chatting amongst themselves half the time.
Like others, I would have hoped for something more along the lines of "This is how you *do* a code review"; in fact, when I mentioned the talk to my colleague, he specifically asked me to try to get this information. Instead, the trio simply ripped into existing code, telling us how this was wrong, that was wrong, and so forth.
I got only two things out of this talk: First, a few useful tools (thanks to Sebastian) for evaluating code; second, the decision to NOT go to the OOP code talk later in the week, simply to avoid the speaker...which is too bad, since it might have been good otherwise.
I really hope Zendcon has a tutorial on real code review in a future year.
Specifically, too much time was wasted with the speakers talking over one another and getting each other off track. Time was also lost deciding which code examples to pull up. You should have come prepared with specific examples.
And like others have mentioned, I expected more in the way of code review strategies or processes that I could bring back home to my team.
I attended other talks given by Sebastian, Stefan and Arne later in the conference so I know they can be good presenters.
Just wanted to clarify my earlier comment (Oct 22, 2009, 18:25). My comment that I avoided Stefan's OOP talk was not really phrased the way I meant it. As written, I seem to suggest that I had a problem with Stefan personally--that's definitely *not* the case. (After all, I don't know him at all, personally or professionally, really!) Instead, I avoided his talk because I felt as though the material that would be covered in the OOP talk would be well over my head--similar to how some of the concepts in the Code Review were over my head.
In essence, I feel that Stefan didn't quite bring the Code Review discussion to a level that was suitable to the audience as a whole, instead catering more to those who were "in the know," so to speak; I was concerned that he would continue this pattern in the OOP talk.
To sum up: I'm hoping to attend one of Stefan's talks in the future, since it's quite obvious that he really, really knows his stuff. However, I stand by my criticisms of the session as a whole (audio problems, lack of direction, mismatch of expectations).
(My new 4-thumb rating is for Sebastian, Stefan, and Arne professionally--these guys obviously know PHP backwards and forwards!)
Comments
Comments are closed.
German comedy at it's best.
Was expecting more of a "How to effectively review code?", not "We will now show you bad code practices".
Good stuff!
Good things - phploc examples and going over user's own code
Bad things - a bit rambling and not enough focus on "how" to review versus doing the actual reviews
Showing how to review code is generally as effective as showing really bad PHP code, I'd suggest limiting it to just one (really bad) example of crappy open source code and then have a checklist for "how to review" and "bad/good practices"
Enjoyable though!
It as great comedy, and fascinating, but I also wish there had been more organization and step-by-step. That said, the three speakers clearly show the tension and compromises that you have make between certain goals (security vs architecture).
The large size of the auditorium as well as the bad quality of the audio system really made this session hard to follow.
Also, it is understandable that the quality of the talk will only be as good as the provided examples, and some of the samples weren't that great (who wants to review good code??)
Though comedic at time, the chemistry between the speakers wasn't great (maybe due to the noise in the room, or problems with the microphones). I think it would have been more effective with just Stefan doing the review, and possibly Sebastian talking about the code metrics.
The first part of the talk was useful (open source projects), but the second part lost my attention.
Terribly disappointing. The sound problems were awful (I could hardly hear anything the speakers were saying) and it didn't help that the trio seemed to be chatting amongst themselves half the time.
Like others, I would have hoped for something more along the lines of "This is how you *do* a code review"; in fact, when I mentioned the talk to my colleague, he specifically asked me to try to get this information. Instead, the trio simply ripped into existing code, telling us how this was wrong, that was wrong, and so forth.
I got only two things out of this talk: First, a few useful tools (thanks to Sebastian) for evaluating code; second, the decision to NOT go to the OOP code talk later in the week, simply to avoid the speaker...which is too bad, since it might have been good otherwise.
I really hope Zendcon has a tutorial on real code review in a future year.
This was a disappointing session.
Specifically, too much time was wasted with the speakers talking over one another and getting each other off track. Time was also lost deciding which code examples to pull up. You should have come prepared with specific examples.
And like others have mentioned, I expected more in the way of code review strategies or processes that I could bring back home to my team.
I attended other talks given by Sebastian, Stefan and Arne later in the conference so I know they can be good presenters.
Just wanted to clarify my earlier comment (Oct 22, 2009, 18:25). My comment that I avoided Stefan's OOP talk was not really phrased the way I meant it. As written, I seem to suggest that I had a problem with Stefan personally--that's definitely *not* the case. (After all, I don't know him at all, personally or professionally, really!) Instead, I avoided his talk because I felt as though the material that would be covered in the OOP talk would be well over my head--similar to how some of the concepts in the Code Review were over my head.
In essence, I feel that Stefan didn't quite bring the Code Review discussion to a level that was suitable to the audience as a whole, instead catering more to those who were "in the know," so to speak; I was concerned that he would continue this pattern in the OOP talk.
To sum up: I'm hoping to attend one of Stefan's talks in the future, since it's quite obvious that he really, really knows his stuff. However, I stand by my criticisms of the session as a whole (audio problems, lack of direction, mismatch of expectations).
(My new 4-thumb rating is for Sebastian, Stefan, and Arne professionally--these guys obviously know PHP backwards and forwards!)