Hacker Newsnew | past | comments | ask | show | jobs | submit | raccer's commentslogin

Would you capture it or just let it slip? Yo


Holy moly, first thread summary I've seen, very cool! Is this an automatic thing now? (assuming url match I suppose)


I used to do this a lot, but I've given up. I used to run a 'bot to do it, but I got a lot of pushback and discontinued.

Dang is a moderator, and he does it when he thinks it's useful. I suspect he does it "by hand" but, like me, has written scripts to help.

Given that your account was created 12 years ago I'm a little surprised that this is the first thread summary you've seen.


Yes - I wrote software that makes it fast to do by hand.

You'd certainly be welcome to resume! I think including the date + number of comments is the important bit, plus not pointing to threads with zero comments - but IIRC you had both of those features.

Edit: one fun fact – it's surprisingly difficult to post links to past threads without coming off as somehow reproaching the latest submitter for posting a duplicate. In reality, HN allows reposts after about a year (and this is in https://news.ycombinator.com/newsfaq.html), but many people don't know that and interpret the list of past threads as an implicit criticism, which is a shame.

I've tried out a bunch of different wordings trying to minimize that misunderstanding. The simplest one seems to be the most effective - I just say "Related". Somehow the word "related" sounds less like a criticism. It also has the nice property of being inclusive, so for example it's fine to include other articles on the same topic.

One of these years, I still want to build software support for collecting related URLs and related past threads into HN's official UI. Then we can all build up sets of related things collaboratively. That will hopefully make HN more interesting.


Then we can all build up sets of related things collaboratively.

You could also take all the posts and related links, toss them in a big adjacency matrix and, oh I dunno, rank the relatedness of pages.

Dumb jokes aside, these do add a lot of extra structure to the dataset and people might think up something interesting to do with that.


This is likely a query issue; I've had many queries go wild & OOM, once fixed they work as you'd expect. There's also a config setting you can use to limit in seconds the amount of time a query can run before it's killed, very useful if you plan on running hand built queries on anything production (which you shouldn't be anyways ;-))


Can you share more info about the queries that "went wild" on GitHub, so that we can reproduce? Thx


Neo4j is awesome, after using at my last startup I'm super meh on table based schemas.


Still use this in examples xD


Not entirely accurate. They have a startup program. Neo4j Enterprise is available for free for startups up to 20 employees.

Source: http://neo4j.com/startup-program/


Ban All the Hammers! PS You're awesome ;)


lol, I wouldn't go so far as calling the villain in any Steven Seagal movie "a wise man".


Seriously, anytime I find a company that freely shares the details of a newer faster way, I'm more a fan. Though with the negative points MS has earned, they're still in the red in my book.


"Day 5 - Embarrassing Flaw #3: Making a stupid business plan, then ignore it anyway"

Awesome! (I lol'd) I love inventing. It's only recently that I've started to add business logic to my imagination, I'm no Jedi in this capacity, but the force was strong at startup school; and I want to train!


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

Search: