Not everyday that you get to see the last 2 years of internal works of the minds of 3 commiters on a major open source web framework. There's some crazy and amazing code in there. Going to take a good while to extract all of the awesome out, but it is bound to be one of the bigger contributions to open source in quite some time.
EveryBlock source code released
11–20 of 22 posts
Re: EveryBlock source code released
#12Earlier quoted context omitted.
> Truly amazing - giving away the source code! They had to as part of the terms for the Knight Foundation Grant. Not that it's any less amazing, or anything. Just saying.
Knight Foundation, as in the people behind KITT?
That was Knight Industries, silly (Knight Industries Two Thousand). Obviously, Knight Foundation is the non-profit arm of the company.
Seriously, though, I assumed it's a non-profit associated with Knight Ridder, who publish newspapers and such. Which I just noticed is only one letter off of Knight Rider. Coincidence?
Re: EveryBlock source code released
#13It's up on Github: http://github.com/brosner/everyblock_code/tree/master
Re: EveryBlock source code released
#14I wonder beyond just city names, what extra data is needed to get an implementation running. What map data is required, how about GPS data? What is the geographic taxonomy... suburbs, zones? I'd like to get something running, but want to have the code as comprehensively implemented as possible.
Re: EveryBlock source code released
#15I wonder beyond just city names, what extra data is needed to get an implementation running. What map data is required, how about GPS data? What is the geographic taxonomy... suburbs, zones? I'd like to get something running, but want to have the code as comprehensively implemented as possible.
I've sent some feedback to EB, but will post here especially with it being now open-source:
--
There should be a page of installations for various countries it has been setup. This may prevent duplicated efforts for various countries, and better promote existing implementations.
Also, non-programmers would be interested ie potential webmasters or administrators. Even a hosted version could be a good business model. As such, check out http://nationbuilder.com when it's released stemming from http://whitehouse2.org
And multi-language? What if I'd like to do something for a single country in two language versions.
--
Also, only Business Reviews, Lost and Found, and Real Estate listings, when clicked bring up actual data. Otherwise, the user sees an information measurement screen before reaching any data. This intermediary step detracts from the site - why can't there be an aggregation of all information from that category in an added feed column that's also available in RSS. This would be especially useful for cities just beginning and when there's not a lot of data, or, when all subcategories are of interest to a user. For example, an rss feed of all police calls for an entire city would be of interest to reporters.
Even Craigslist has an aggregation, per city > per information top-category: http://telaviv.craigslist.org/jjj/ .. with rss
Re: EveryBlock source code released
#16Unfortunately, the HTML and CSS that makes EveryBlock so useful is not included in this release.
I'm not sure why the Knight Foundation didn't insist that these components be part of the release, unless the designers were paid out of a separate budget.
I understand that releasing these components would have led to clones of the site, but I think, to a certain extent, that is one of the goals of the project.
Re: EveryBlock source code released
#17Unfortunately, the HTML and CSS that makes EveryBlock so useful is not included in this release. I'm not sure why the Knight Foundation didn't insist that these components be part of the release, unless the designers were paid out of a separate budget. I understand that releasing these components would have led to clones of the site, but I think, to a certain extent, that is one of the goals of the project.
HTML/CSS is the easy part. Screen scraping and normalizing/organizing all the data and writing a custom mapping service were the tough ones.
Re: EveryBlock source code released
#18Unfortunately, the HTML and CSS that makes EveryBlock so useful is not included in this release. I'm not sure why the Knight Foundation didn't insist that these components be part of the release, unless the designers were paid out of a separate budget. I understand that releasing these components would have led to clones of the site, but I think, to a certain extent, that is one of the goals of the project.
HTML/CSS is the easy part. Screen scraping and normalizing/organizing all the data and writing a custom mapping service were the tough ones.
I beg to differ. Coming up with an easy-to-understand, easy-to-use interface for the surfeit of data is incredibly difficult.
Re: EveryBlock source code released
#19Earlier quoted context omitted.
HTML/CSS is the easy part. Screen scraping and normalizing/organizing all the data and writing a custom mapping service were the tough ones.
I beg to differ. Coming up with an easy-to-understand, easy-to-use interface for the surfeit of data is incredibly difficult.
EveryBlock could have been successful with any number of designs, the interesting part about the site is the data and its organization. I agree that it's great looking, but a site doesn't have to be great looking to be a hit (Craigslist for example).
Re: EveryBlock source code released
#20Earlier quoted context omitted.
I beg to differ. Coming up with an easy-to-understand, easy-to-use interface for the surfeit of data is incredibly difficult.
EveryBlock could have been successful with any number of designs, the interesting part about the site is the data and its organization. I agree that it's great looking, but a site doesn't have to be great looking to be a hit (Craigslist for example).
True, it could have been.
To me, it would have been nice to have the design to use as a starting point.
It also seems that it's part of the source code of the site.
When reddit went open source, they included their CSS, images, HTML, etc. I think this makes it easier to get going. Reddit and EveryBlock are going to continue to innovate enough to differentiate themselves from the clones.
I also hope they've included the mapping code/JavaScript. I haven't had a chance to look through the source to see if this is the case.