This is great, Peter, thank you for this perspective!
If you walked away with the impression that there were gotchas, then we have failed. While we do have specific criteria we're looking for (as we're trying to make the grading of each work sample as objective and consistent as possible), I wouldn't call those items "gotchas". It's also not the type of thing where if you miss one small item, you're disqualified. The grading allows for several missed items and is only looking for some total of all allowed points (across 33 different criteria). So while some of the items you didn't get credit for might have appeared trivial, or like a gotcha, there should be enough allowance in the process to not let one or two misses scuttle the whole effort.
> Perfect candidates do not exist; just as perfect humans do not exist. Great test takers do exist though. Personally, I have never correlated good test takers with good employees or good programmers
We are totally with you re: perfect candidates. Work samples are only one part of the hiring process after all, and are meant to ensure some base level of technical ability. Others parts of the process, such as our structured interviews, are meant to flesh out some of the nuances I believe you're referring to here. We're not looking for a perfect candidate, but we are looking to understand what makes up each candidate to know if their strengths are a fit for Spreedly and if we can live with the weaknesses (and vice versa for the candidate evaluating Spreedly!).
We really do appreciate your effort, both in writing up your thoughts here, and in applying in the first place. This is how we get better at this kinda of thing!
I work on Heroku, on the databases, which are rather thorny in that department...
...it's going to take some evolution in how applications are written to make that work not uglifically (and it'll probably be ugly at first and for a while), but it's the dream, and I think it'll come to pass over time and with work from many people.
There are literally more Heroku folks in this thread than unaffiliated commenters :P well played.
Personally I'd just be happy with a master-master postgres db or sharding, for $money, where money is some amount enough to make that happen without me worrying about it.
The lack of search is the main reason I created http://gistedapp.herokuapp.com. Unfortunately I still find their search lacking in a few fundamental ways:
* it is (surprisingly?) case sensitive
* it only searches titles, not file contents
* the UI feels disconnected from the main gist area, making it unclear whether its a search of all Gists (like the previous version) or a personal search.
Fortunately, this update indicates they're paying attention to a much-used part of their product that was starting to feel neglected. At their pace of improvement I can't see these holes remaining for too long.
Now, hope they add back the ability to be notified when there's a comment on your gist (which Gisted supports)!
Just read your post, njyx. While it's solid and thorough, my first impression is that it's a bit too high-level for an audience of application developers.
This is great, Peter, thank you for this perspective!
If you walked away with the impression that there were gotchas, then we have failed. While we do have specific criteria we're looking for (as we're trying to make the grading of each work sample as objective and consistent as possible), I wouldn't call those items "gotchas". It's also not the type of thing where if you miss one small item, you're disqualified. The grading allows for several missed items and is only looking for some total of all allowed points (across 33 different criteria). So while some of the items you didn't get credit for might have appeared trivial, or like a gotcha, there should be enough allowance in the process to not let one or two misses scuttle the whole effort.
> Perfect candidates do not exist; just as perfect humans do not exist. Great test takers do exist though. Personally, I have never correlated good test takers with good employees or good programmers
We are totally with you re: perfect candidates. Work samples are only one part of the hiring process after all, and are meant to ensure some base level of technical ability. Others parts of the process, such as our structured interviews, are meant to flesh out some of the nuances I believe you're referring to here. We're not looking for a perfect candidate, but we are looking to understand what makes up each candidate to know if their strengths are a fit for Spreedly and if we can live with the weaknesses (and vice versa for the candidate evaluating Spreedly!).
We really do appreciate your effort, both in writing up your thoughts here, and in applying in the first place. This is how we get better at this kinda of thing!