Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Show HN: Recalll – Search Engine for programmers and developers (recalll.co)
69 points by parvbhullar on May 28, 2015 | hide | past | favorite | 39 comments


It reminds me of an article from Paul Graham: http://www.paulgraham.com/ambitious.html

"The way to win here is to build the search engine all the hackers use. A search engine whose users consisted of the top 10,000 hackers and no one else would be in a very powerful position despite its small size, just as Google was when it was that search engine. And for the first time in over a decade the idea of switching seems thinkable to me.

Since anyone capable of starting this company is one of those 10,000 hackers, the route is at least straightforward: make the search engine you yourself want. Feel free to make it excessively hackerish. Make it really good for code search, for example. Would you like search queries to be Turing complete? Anything that gets you those 10,000 users is ipso facto good."


hahaha man, you made my day :) thanks


Great project, a usable search engine for programmers would be a fantastic tool to have.

Realising that Google still delivers the most relevant results for the things I want to look up quickly I created http://www.stacksnippet.com/ last year. It takes the first Google results from your query and displays stack overflow answers inline.

Recall seems exciting, I'll be sure to follow it.


thanks alot, stacksnippet.com is also a good tool.


This is search engine for developers and programmers, it enables you to search across websites as well as cloud apps. I'm looking for feedback. Thanks


Cool efforts so far. I can see the use case for wanting to find information across multiple sites.

Couple feedback points.

1 - the ui in your screenshot on the home page (https://recalll.co/theme/flat/img/search.png), in the results section you have the title as blue. At least in safari, in real life its black. This makes the results all run together and visually be hard to quickly tell where one result ends and the next begins.

2 - if I was to use it for actual work, the search "nsfetchedresultscontroller duplicates" would be something that I have searched for recently. unfortunately nothing really useful came up on this recalll.co. Maybe a good example to test against google, stack overflow, and quora.

3 - not sure if the types of queries I want to put in to search asana or evernote, would be the same types of queries I would want to search the web / blogs / stack overflow. I suppose maybe the use case is, just use recall, and we'll figure it out. But sometimes I may want to search for something silly like "s3 credentials from bob", which might producer weird results in stack overflow for example, because there is probably lot of people talking about s3 credentials, but its completely out of context, and I would want the results from evernote or something. (whether evernote is a good place to store those of course is probably another discussion, but to illustrate the point)

4 - some people won't install chrome extensions. if I try to search right from your homepage, the animated help thing starts to overwrite my input. whatever I'm searching, if its not already open, I want to open it quickly, and execute the search immediately.

Great work though. The concept is cool. Maybe you could end up like duckduckgo, for programmers.


thank man, for your valuable points.

For 1 - Ya I also think blue is good for title color. i changed it some time back to black will revert back to blue.

2 - due to lack of data results are not good, crawler is continuously running, once we have enough data, that wont be a problem.

3 - for cloud apps - you hit exactly on the point that's why we added cloud apps, to target specific site or app you can choose app name in query like "s3 credentials from bob evernote", it will do the trick for you until more usable is developed.

4 - Yes chrome extension installation is bit tricky but right now don't have anything more innovative will figure it out. Search form on home page is symbolic, its not actual search. Initial idea is to take user to search page, definitely club both the pages to single one, like most of search engines are doing.


Cool, keep hacking away. Bookmarked.


Found something interesting.

If I use it like it is google and search for "execute javascript capybara" I get awful results.

If I use it like I am talking to someone and say "How to execute javascript with capybara" the first result gives me exactly what I want.


Reason behind that we give priority to phrase based searches over terms based searching. though we are figuring the way for more relevant searches for any kind of query.


Searched for "javascript date", got the date. Didn't know what to expect. "Thu May 28 2015 16:41:46 GMT+0100 (WEST)"

"javascript while(1)" also gives interesting results


there is broken code while searching for particular queries. will fix that. thanks for pointing out.


really interesting


I'm happy to finally find a search engine giving good results for `>=> haskell`[0] as I can never find good information on those symbols and never remember them.

One problem, the link to the browser extension seems to be broken [1].

[0] https://recalll.co/app/?q=%3E%3D%3E%20haskell

[1] https://chrome.google.com/webstore/detail/recalll/kimdaoebjj...


For Haskell in particular, you are better off using Hoogle [1] or Hayoo [2].

[1]: https://www.haskell.org/hoogle/?hoogle=%3E%3D%3E [2]: http://hayoo.fh-wedel.de/?query=%3E%3D%3E


You're right, Hoogle and Hayoo are pretty great (search for a function by it's type signature!), but I still think it can be helpful to see other sources of information. Good point though.


good to hear that, it answered your query. yes chrome extension link is broken, fix will be available soon.


Seems like it's fixed now. Thanks.


Nice idea, nicely executed. Tiny suggestion: I'd put a little more white space between the search results. I'm finding it hard to distinguish between different results.


sure.


I was playing with async controllers in .NET last night and was looking up a lot of reference material. Recalll's results were surprisingly great- well done!

https://recalll.co/app/?q=.net%20async%20controller


glad to hear that. thanks


I love the idea, so I immediately gave it a shot.

https://recalll.co/app/?q=haskell%20fibonnaci doesn't give me any useful results. Same query on google gives me exclusively useful results (at least on page 1).

I wish you the best for your project, and I'm sharing my use case in the hopes you can continue to improve, not to discourage.

Edit: Doh; it's because I can't spell. Two c's and 1 n and I do get some relevant results. Google has the "fix it for me" hidden from me.


will add query suggestion and "did you mean" functionality for more suggestive queries.


Searching for "racket parallelism" does not really give anything relevant. In this case I searched for something that is extensively written about in the actual Racket documentation, both in the guide and reference material, so I know what I would expect.

It's a very natural search to do if you're thinking about parallelism and concurrency (both of which are suggested by Google, for example), but I assume most Racket users use the excellent documentation search for Racket. If you indexed that you'd probably get great results.


yes due to lack of data results are not relevent, definatally index documentation for racket. good points though. thanks


I asked it what is Lucee? I got back some gibberish about JavaScript. The answer is an open source scripting language on the JVM.

The first ten replies on Google were correct as was Wikipedia: http://en.wikipedia.org/wiki/Lucee

Wondering just when you did your crawl as Lucee has only been around since January?


the reason is we didn't crawl wikipedia yet, we are planing to do that very soon.


Google still finds way better results for “complex gamma function”, for example. I have yet to see a search engine that can take on the Goog, let alone surpass it. I wonder what it would take to accomplish that...


What you are saying is ambitious, this is first beta release of product, we do have vision to take on google at least for programmers but it will take time, right people and resources. I hope next release much better than this one in terms of speed, relevancy, technology, search algos and data. though very motivating points. thanks


I realise it's ambitious, but in the end, if you can't at least (have hopes to) compete with Google, what would be the point in starting a search engine project at all? Roughly, all your prospective users are Google users now.

You have my best wishes to succeed.


totally agree with you. as i told you we have vision to acquire users from mainstream search engines, what i'm asking is more time. though thanks a lot for your wishes.


this looks interesting; though i think the design should be more search focused like how google's homepage is; with the searchbar at the top is good for results, but it should be more center for the homepage; like, there's a bunch of other stuff going on there that I don't care about; i'm there to search. It's distracting, really.

Otherwise, keep at it! the more diversity we have the better.


good points, added to TODO.


Pretty neat idea! I think it would be really cool to declare my default languages though, so it only returns results from what I specified.


good suggestion, added to TODO.


Great idea, will use.

One thing I'd really like to see is cheat sheets. For example, "regex cheatsheet".


great suggestion.


Or..

$ pip install howdoi

$ howdoi reverse string python

> 'hello world'[::-1]

> 'dlrow olleh'




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: