Re:Seriously, we need multi-column editors
by lowbloodsugar on Wednesday March 04, 2015 @11:35PM
(#9149274)
If your class is so big that you need to have multiple views of it, then it probably doesn't have great SOLID metrics.
Re:Seriously, we need multi-column editors
by paperwork on Thursday March 05, 2015 @01:23AM
(#9149544)
You don't have any files over 200 or so lines of code?
Re:Seriously, we need multi-column editors
by lowbloodsugar on Thursday March 05, 2015 @04:15AM
(#9149858)
I do and I regard them as a bug in need of refactoring. I just ran sloccount on one of our key services and the largest non-test file is 106 lines. It has three methods that do not call each other. The largest test file is 268 lines.
Then again we have another service with four files over 1000 lines (1441 the biggest), and we experience significant pain: not because of the size of the files but because of what the size of the class means about its design. That's on a team that had significant pressure to meet some tight deadlines. They are now actively working on refactoring those classes - again, not because of the size, but because of the design.
Just curious, why is emacs not an acceptable solution?
Re:Seriously, we need multi-column editors
by paperwork on Wednesday March 04, 2015 @12:09PM
(#9145667)
Honestly, just never used it (not since freshman year almost two decades ago). I mainly use eclipse and it works very well for me. I like how I can hover over a variable and it tells me info about that variable, click on it and it shows me very other place it is used. I can rename a function and it gets updated in all files where it is used. I can very quickly see where a class fits in the class hierarchy or see which methods call other methods.
Emacs might do all of this, but eclipse does this well enough.
Regarding the anti-pattern. You may be right, I have't updated the code in a few years. I displayed it in the screenshot only because it is already open source. But I do need the option of writing files to disk, writing to some random data store or just keep the data in memory. Send me a note about your idea, the whole point of opening it up was to get feedback.
QuickFIX is pretty good, but doesn't hurt to experiment :)