Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Mwmbl: Free, open-source and non-profit search engine

mwmbl.org

Post reply on HN
Print view View original

101–110 of 129 posts

Re: Mwmbl: Free, open-source and non-profit search engine

#101

Post by brettermeier »

arghwhat wrote: post #65
brettermeier wrote: post #47

Earlier quoted context omitted.

But it's a normal word, unlike "mwmbl" (I had to look it up, couldn't remember where the "b" and "w" goes after some seconds).

No it's not, it is an intentional misspelling of "googol" and means nothing - not in English or any other language. "Spotify" is also not a "normal word" in any language. And for those not native in English (there's supposedly only some 400 million of those), it's just a random sound sequence like any other. mwmbl is a shortening of the welsh writing of https://en.wikipedia.org/wiki/Mumbles . Only tricky part is know…

Oh thanks for clarification.

Re: Mwmbl: Free, open-source and non-profit search engine

#102

Post by daoudc »

marginalia_nu wrote: post #26

I was happy to notice these guys a while back, but the git repo seems very dormant. I wonder if they backpedaled on the open source side of things, or if the project is asleep. Either would be sad, because the world needs more open source search engines.

Yes, I took a break for a while, my fourth child was just born! Still committed to the project though and working on it when I get time.

Thanks for your encouragement. Would love to have a chat some time.

Re: Mwmbl: Free, open-source and non-profit search engine

#104

Post by marginalia_nu »

daoudc wrote: post #100
marginalia_nu wrote: post #69

Earlier quoted context omitted.

Yeah. That seems like a design decision that will scale poorly. For reference, even in my dinky 100M index I have individual terms with several gigabytes of associated document references. In general hash map table index designs don't tend to be very efficient. If you use a skip list or something similar, you can calculate the intersection between sets in sublinear time.

We actually just take the union and then re-rank. Because the lists are all small, this is cheap.

Point is, with a skip list (or similar), the lists don't need to be small. You can intersect data sets that are enormous very quickly using this algo[1] where a single linear read of both lists is the worst case scenario.

[1] https://nlp.stanford.edu/IR-book/html/htmledition/faster-pos...

Re: Mwmbl: Free, open-source and non-profit search engine

#105

Post by marginalia_nu »

daoudc wrote: post #102
marginalia_nu wrote: post #26

I was happy to notice these guys a while back, but the git repo seems very dormant. I wonder if they backpedaled on the open source side of things, or if the project is asleep. Either would be sad, because the world needs more open source search engines.

Yes, I took a break for a while, my fourth child was just born! Still committed to the project though and working on it when I get time. Thanks for your encouragement. Would love to have a chat some time.

Feel free to shoot me an email, I love to talk search engines :D

Re: Mwmbl: Free, open-source and non-profit search engine

#106

Post by tamimio »

marginalia_nu wrote: post #90
tamimio wrote: post #79

I searched first test “best business banks in Canada” and it showed no results saying it couldn’t find any “We could not find anything for your search..”, I can also see two redundant lenses icons.

What sort of result would you expect for such a query?

Primarily a list of options to choose from, preferably that from a non-affiliated site, asking the same in GPT-4 I get the following:

>Tangerine Business Savings Account: This account offers a high interest rate of 2.65% to 3.25% on your balance, no monthly fees, no minimum balance requirement, unlimited transactions, free e-transfers, and access to over 3,000 ATMs.

>Wise Business Account: This account offers low-cost international payments in over 50 currencies, no monthly fees, no minimum balance requirement, free local transfers, free debit card, and access to over 10 million ATMs.

>BMO eBusiness Plan: This account offers no monthly fees, no minimum balance requirement, unlimited transactions, free e-transfers, free cheque deposits, and access to over 3,500 ATMs.

>RBC Digital Choice Business Account: This account offers no monthly fees for the first three months ($5 per month thereafter), unlimited electronic transactions, 10 free debit transactions per month ($1.25 each thereafter), free e-transfers, free cheque deposits, and access to over 4,200 ATMs.

Re: Mwmbl: Free, open-source and non-profit search engine

#108

Post by luc_rnz »

Thank you for sharing this, this is very interesting. I will give it a try, although I don't think it can replace my current engine (DuckDuckGo/Searx), but rahter complement it maybe (by having a smaller, more curated set of data).

Particularly I am having a great time reading the crawler extension source-code: https://github.com/mwmbl/crawler-extension

Re: Mwmbl: Free, open-source and non-profit search engine

#109

Post by marginalia_nu »

tamimio wrote: post #106
marginalia_nu wrote: post #90

Earlier quoted context omitted.

What sort of result would you expect for such a query?

Primarily a list of options to choose from, preferably that from a non-affiliated site, asking the same in GPT-4 I get the following: >Tangerine Business Savings Account: This account offers a high interest rate of 2.65% to 3.25% on your balance, no monthly fees, no minimum balance requirement, unlimited transactions, free e-transfers, and access to over 3,000 ATMs. >Wise Business Account: This account offers low-cos…

Well if we imagine a search engine as a document retrieval machine, who would publish such a document?

Re: Mwmbl: Free, open-source and non-profit search engine

#110

Post by Brian_K_White »

eviks wrote: post #31
mdaniel wrote: post #15

I wondered if this approach would be feasible for a distributed crawler: https://github.com/mwmbl/mwmbl#crawling Also, your own posting appears to be missing from the index: https://mwmbl.org/?q=mwmbl+ycombinator (and, yes, another vote for changing the domain name; you can have a quirky project name, but if I can't remember the cat-walking-on-keyboard domain, I'm not going to use it)

You don't need to remember it, just bookmark and tag however you like (it's anyway a waste of keystrokes to manual type the full domain for such a frequently used site like a search engine)

This is 100% wrong.

One's normal phone and laptop is actually only a fraction of uses, and even one's normal device isn't just one thing that needs to be done one time in grade school and then set for life. It's a dozen different things, and they are all perpetually rotating, and most people are not highly optimized with profiles they actually export and import.

This idea is great but it's going absolutely nowhere without a better understanding of actual humans.

Post reply on HN