Prompted by this post, I wanted to take a moment to encourage people to try using the built in SuttaCentral Search again. Itâs greatly improved from even a few months ago and is really worth your attention.
I know it was really bad in the past and I too had given up on it for most things. However itâs working well now and is always the first thing (and often the only thing) I try.
There is even a hidden feature that allows you to find all translations by a specific translator. Just search for author:suddhaso for example.
But it wonât continue to improve unless people start using it and letting us know what works and what doesnât. There are still problems and lots of ways it could be improved. But it really is worth your time to give it a try.
One cool thing you might not have noticed in the current search is that the âbreadcrumbsâ are given for each sutta and they are actually clickable. For example, I did this search for turtle. This was one of the results:
One thing I am noticing, the instant search is no longer instant. Like, I type in mn1 I should get a list of all suttas with that in them before I type the next number, then it should filter down.
It used to do this, but now I type in mn1, nothing happens, I then type the next number mn18 and it waits a second then shows me results. Itâs not a cache or network thing, it happens persistently. I literally type the ID, then count âone thousand and oneâ and then the results appear. Itâs too slow, it should be an imperceptible lag, thatâs why itâs called âinstantâ.
Not sure why this is so, but we need to fix it. My guess would be that either a bug has crept in or that complexity or too much data is causing the delay. If the latter, then we should focus on getting the relevant results instantly and lazy-loading other things.
Is anyone else having the same problem?
Relatedly, when the results do appear, the overlay janks: this is a CSS bug. It needs to define the absolute height of elements first so they donât create jank when appearing.
Iâd look into fixing this myself, or at least making a proper issue for it, but iâm traveling and apparently locked out of Github!
You can see that now the first thing to click will run a full database search restricting it to just early Buddhist Suttas (So EBT minus Vinaya). All the other items will take you right to a sutta.
If you just press enter (or the search icon at the end of the line) it will run a full database search, which looks like this:
Thatâs not related to search, though, is it? If you are on mobile and just need to read a sutta, you might try my SC Light app. It loads instantly. Itâs using the same API as the SC real site. But of course it doesnât have the lovely search Iâm trying to promote in this thread.
If you are having problems with the site in general you could start a new thread to get tech support.
At the moment, all Pali diacritcs are ignored for search.
If anyone finds that diacritics are not being ignored, please report it in the search feature bug page ASAP. Please give exact searches with screen shots if possible so we can figure out what is happening.
Hmm ⊠I didnât actually test it on SC. I just assumed from the fact that @BethL didnât get the expected results that this might have been a reason.
Found the other topic about the same issue, deleted the offline app for sutta central which is not working anyway, cleared 4 GB of data stored in Microsoft edge, now itâs fast.
To clarify⊠Speaking in terms of the Latin alphabet⊠The letters are not being ignored. Letters with and without diacritics are treated as equal. So for example, a will match both a and Ä. This allows the vast majority of people who do not have Pali typing on their device to run searches.
So if you search for sammaditthika it will find sammÄdiáčáčhikÄ. In the entire Pali canon as well as commentaries and sub commentaries, the term sammaditthika does not exist so there is not going to be any confusion.
I have so far not run into cases where ignoring diacritics has caused problems. But then again, I never use SC for Pali searching. The DPR will always be far superior for that kind of search.
Could you talk a little bit how the feature of ignoring is causing problems? If we can find enough cases of problems then perhaps we could add a setting to turn the diacritic matching on and off.
Granted. ⊠I see what you mean from the mechanical search pov.
Itâs just that thinking of Pali phonemes in terms of Roman letters scrambles my learnerâs mind. I find it extremely difficult to learn the correct pronunciation and spelling of Pali and doing anything at all that permits me to ignore the correct sound-symbol relationship seems a hindrance. ⊠Because my learning is slow I welcome anything that forces me to pay close attention to how the language is written,
[The Latin alphabet is such a mess when used for modern languages, something that I struggled with in elementary school and undergrad. Then I went to India and was introduced to the Bengali script, a close cousin of the DevanÄgarÄ« with its symbols following a phonetic order. I was completely bowled over and became attached ⊠There is such beauty in this whole concept ⊠:D]
This is an important consideration, and one I didnât stop to consider.
Thank you for the explanation. I wish I had been more considerate and not made that remark.
So to summarize⊠So far this behaivour in search is not actually preventing you from finding the things you want? Itâs that it throws up a roadblock of your mental map of the Pali alphabet concept?
I fully admit that natively t and áč are no more similar than k and c are. However it is unlikely that a user would switch or mistake the latter but almost a guarantee they would mix up the former (or not be able to type one of them).
If I had my way we would also ignore aspirations and double consonants since they are also a common typo people make. In fact if you have ever used my sutta name lookup tool you will see that there is an optional fuzzy search that ignores not only diacritics but also aspirations and double consonants. You would think that this would throw the world into chaos, but I only ever use the fuzzy search and I find what I want every time.
Of course once you get into such drastic fuzziness you may end up catching a lot more than you want. If something like that was ever implemented in SC search it would probably have to be an option that could be turned on or off.
In the instant search we are (I believe) making use of a built in misspelling tolerance, but you will notice if you pay attention to the results at the end of the list that it returns wildly implausible suggestions.