Hacker News Reader

Comment by SpeakMouthWords | Hacker News Reader

The writer appears to be the wrong audience of !bang. As a physics student, I find !w tremendously helpful. Image searching is also much faster since it all can be done in the search bar rather than requiring the mouse. !wa for wolfram alpha is again great. It's far better for sites in which there is no better alternative than using their search bar, rather than the type of usage the article seemed to employ.

View on HN

You can set up this sort of thing at the browser level to work directly with the site of interest (Wikipedia in your case):

http://support.mozilla.org/en-US/kb/Smart%20keywords

You can set up the same sort of thing in Chrome by going to Settings -> Manage search engines, then editing the middle column to have just one or two letters of your choice.

Keyword bookmarking is also very useful: http://lifehacker.com/196779/hack-attack-firefox-and-the-art...

In fact, at this point, the number of times I have to manually type in a URL (when using my own computer) is probably fewer than 10 per day. I've never understood why people make such a big deal about !bang when instead you can search the sites directly with a keyword search, plus you're not limited to those DDG supports.

You can manage your search engines in Chrome to get the same !bang effect. Slightly better in my opinion, because you don't have to type the awkward '!' and you don't have to go to a website just to search another website.

For example, if I have the URL http://www.wolframalpha.com/input/?i=%s tied to the 'wa' keyword, I can now search Wolfram Alpha from the address bar by typing 'wa'.

The best thing is you can even configure Chrome to do the same thing for local files. My Racket repository is always in ~/code/racket. The "rkg" keyword (for racket-git) is set to file:///home/andy/code/racket/doc/search/index.html?q=%s, which allows me to easily search all of the latest racket Git docs in Chrome in just three letters + Tab.

Thats only because the search for wikipedia is good; but for example !python isn't as great. It is nice for convenience but relying on site's internal search is kind of a cop out, albeit smart for their purposes.

Ok, !python, in particular, sucks. So don't ! the python. Everything else is a plus though. !hoogle is awesome.

If you can't use it for all sites you want to search, then what's the point? I'm sure python isn't the only site with a bad search.

I really have to ask if your point is that because a feature isn't universally useful it's automatically pointless?

I think that the approach that DDG is taking is wrong. Relying upon individual sites search engines isn't going to lead to uniform results because they are always going to vary in quality.

But look at a query like "!hoogle (a->b) -> [a] -> [b]". That's a search on the type of a function, and I'd have to go to hoogle first, then search it from there. This adds value greater than the drawback of not using !python, or hitting the back button and deleting the ! when I forget

I don't see that they realistically have an alternative.

DuckDuckGo won't ever be able to compete with Google as a general search engine. Google already pretty much solved that problem, and even if they haven't they have a huge head start and a much bigger bankroll. If DuckDuckGo is going to succeed on it's search principles it's going to be as a curated search engine. That's a game where they can still compete with Google.

I completely agree with you about the lack of uniformity in search results. Curation is hard. It requires that a good source exist, that the curators are aware of it, and that it's amenable to being added to your system (technically and legally). Only one of these is under the direct control of the curators.

The reason the Python search sucks in DuckDuckGo the people running the curated Python search have a crappy search engine. Obviously either a better choice doesn't exist (probable), the DuckDuckGo curators don't know about it (very likely), or the better source isn't technically or legally accessible to the DDG curators (unlikely but possible).

Unlike the problems of a crawling and building general search index (where even to enter the market you need to have a massive infrastructure investment), these smaller problems of curation are attackable by a smaller team with limited resources. To improve the search results for !python they won't need to re-engineer their entire crawl/index infrastructure, they will simply need to pick a better third-party source once they are aware of it existing.

It just seems he used a useless search on purpose. The !w for wikipedia searches and the !v for youtube searches are as awesome as the !python is lame.

Or he is a Python developer(I know I work with him). Searching the python docs or the Internet for some module is a pretty common use case.

True, but I think DDG is smart to pursue the bigger market of "people who want to find stuff on wikipedia, youtube, and google image search" and wait for DuckDuckHack to come up with a plugin that does skillful python searches.

Should plugins really be necessary? What would a plugin bring to the table that searching "python whatever" doesn't?

Reply on news.ycombinator.com