SC Voice – the road to v1

Oh wow! That is a new one for me. I have added first point.:white_check_mark:

Also, regarding the earth, water, fire, and air individual search terms, each of these search terms returns over 50 suttas. The terms are used all over. In contrast, if we search for them together as earth, water, fire and air we find a more focused set of 10 suttas. Are there element-specific searches in the suttas? I have not found any and they seem to always be used interchangeably. If we cannot find element specific uses, then I would suggest that we replace the four individual element search terms with the single phrase earth, water, fire, and air.

2 Likes

I’m sure you’re perfectly aware that I’m just honorarily sitting in as PM for camaraderie and you’ve got a good sense of what you want to do with your project which you’ve stewarded beautifully to this point.

For what it’s worth, I think if you’ve meaningful fixes ready to go, it makes a lot of sense to release them as releasing itself is pretty trivial, so yes, v0.9.2 sounds like a good idea. Likewise, if you have a sense of drift, I’d probably encouraging focusing on the core value you want to deliver for v1.

Hope this helps. :slight_smile:

2 Likes

Yes it is. Would you mind making a surgical split as you see fit in the Release Plan?

2 Likes

Okay, I’ve done some very quick field surgery to try and help focus down on what you’ve raised above. Therein the suggestion is that v0.9.2 exclusively deals with pronunciation issues. Thereafter, performance and little usability fixes polishing for v1* and after that perhaps regroup and see what’s still relevant on the list and what would add most value for users.

Again, I stress this is just suggestion trying to help reflect back what I’ve taken from the above. If you think it would be best to invest your time in another way… please do rush the RP back into theatre. :wink:


* out of curiosity with respect to downloading search results as a single file was this a request from someone? I don’t quite understand it’s appeal. I completely get the wish to download all search results individually, but not as a single file… might just be me though.

2 Likes

This is certainly a good one! It doesn’t cover all the water and fire similes I had in mind when proposing the individual words, but you’re quite right that these words return too many results.

2 Likes

The need to download MP3’s arose out of my own need to listen to suttas as I walk meditation. Each day I spend about 90m out there. Walking out there I have no internet connection and the MP3’s are the only solution. And when someone suggested a sutta play list, I immediately thought I wanted that as well. For example I want to play MN44 with SN12.23 but I cannot today, so I will need to wait till January to walk the streets listening to a Pali/English playlist.

But the wait will be worth it. I think it is important to get the translation right. The words really do seep deep through repeated listening. And I think we would all want the right pronunciations to seep in.

4 Likes

Very true! I’m also very much in favour of downloadable MP3’s. But to me individual files would be just as good as files with multiple suttas—even better though!

3 Likes

Yes, as I said, I completely understand the value of download. It was the single file aspect that I was asking about. As someone who listens to downloads, (as a general point not related to SCV) I’d personally prefer to have two separate files (eg MN44 and SN12.23 using your example) rather than one file - my (and I think pretty much every) audio player has a playlist feature, so this is why I was interested to know the benefit of one single file.

2 Likes

Just checking search phrase “study”: It returns two results, Thig 9.1 (which contains the keyword), and SN 55.53—and in the latter I can’t see the word “study” anywhere.

This may be a similar problem to what happens with the keyword Sabbamitta which returns some results that don’t feature the word.

It would be quite a job to systematically check them all!! :open_mouth:

1 Like

Ah. OK. Now I understand. Perhaps it is my own lack of skill that is the barrier. I am playing the single 90m file on DropBox. I do not know how to listen to the individual audio sutta tracks on my iPhone because Apple’s desire for monetization has made that one task extremely difficult to me. I’m afraid to hook my phone up to a desktop because it screws up my phone all the time. So I learned how to use Dropbox and that requires a single file playlist.

Also, unlike music, which I happily shuffle, I find that walking the suttas mindfully works best for me following exactly the same invariant path through the neighborhood. This 90m of walking paces out the sutta experientially so that at a particular tree I am hearing a particular text segment. It is a powerful mnemonic aid. Shuffle would destroy that value. So I carefully create single file playlists of suttas in a sequence I deem useful for practice. And then I glue them together as one file for uninterrupted listening.

Judging from your feedback and Sabbamitta’s, I’m thinking my perspective is idiosyncratic and hence of lower priority.

Added tov0.9.2 Release Plan.:white_check_mark: Thank you!

2 Likes

Right, now I can really see (and from other contexts of trying to wrestle with getting things to work with my setup, hugely empathise with) the issues your trying to solve and I think you’re spot on, it’s that tricky issue of speaking to specific user cases.

In terms of what you’re trying to achieve I’d aim for exactly the same thing 90m of material in a specific sequence, it’s just that the particular sticky points you’re negotiating haven’t ever come up for me - I don’t use shuffle and I don’t use Apple. The main thing I was thinking when asking after the point was just trying to understand it’s value.

The only real advantage I had in mind in terms of (my personal preference for) individual tracks is that they make navigation easier. If a user has a slightly different scenario than a highly fixed 90m walk were perhaps they might have cause to pause, they’d be able to go to the beginning of the previous sutta. Or, horror of horrors, if eg. they think “hey, this sutta isn’t really what I need to listen to right now” they can skip to the next one (I really admire your listening principle of just taking what comes, but equally see the benefit of helping users to find the right way for them at a given time).

I think we’ve both said that the ability download a playlist is a high priority. In terms of whether having them as one file is a high priority, I don’t think a sample of two is big enough to assess, but I would agree that building to one potentially not very common use case isn’t ideal. My answer: ask more people! :grin:

1 Like

Yes. I agree. And the problem of split files even happens with my own use case. For example, I would like to listen to DN33 in Pali/English. However at 5 hours, that would be too long for a single file. So the need is actually for downloading a playlist of multiple files. And this is the tricky part. Some suttas are short and easily concatenatable. Some suttas are long and have to be split to achieve multi-file download. This is the difficult part of v1.0. And perhaps we should have a v1.1 with the ability to split long suttas into sections.

For technical reasons, the following are distinct problems to solve:

  1. Single file download of playlist.
  2. Multi-file download of playlist split by sutta. This is a convenient step up from doing #1 one by one. But it doesn’t solve my problem.
  3. Mutli-file download of playlist split by sutta and/or section. This handles DN33 split and has no other solution.

Multi-file download is tricky because it replaces the simplicity of a single file with all the complexity of multiple files. I.e., what if file1 and file 3 downloaded but not file2? With single files, we simply try again until we get what we want.

2 Likes

Yup, really see the conundrum.

Again, first and foremost, I’d encourage you to follow your instincts in terms of what you want to put your energy into. Thereafter, I’d probably suggest following the simplest avenue that will allow you to get to v1 and then build on top of that (naturally, a major caveat being that if a ‘simple’ decision now will inhibit development later then it’s worth pausing to find the better solution now).

As a complete wildcard thought… one of the things that popped into my mind as I looked at the RP yesterday was whether how SCV is built would work for something like Downthem All (then I paused to mourn the loss of Downthem All - the only way I’ve been able to get this functionality post FF57 (? I think) was to switch to Chrome and the Chrono Download Manager extension)…?

2 Likes

Based on our discussion, I’ve updated the Release Plan to have v1.1 download multiple files. This is the harder problem to solve. the v1.0 single file download will provide a single file playlist of one or more suttas in Pali, Pali/English or English according to preference. In this way I think we chip away at the problem eventually solving for all.

I am an old fuddy duddy not inclined to explore new-fangled technology unless prompted with fiercely wielded cattle prods by vigilant passionate people on a mission to change the world. This is why I am a lousy PO. If left to my own devices, I would simply solve for me and ignore all. Because of this, I would welcome any guidance you might assert resulting from your own explorations and conversations with others.

2 Likes

Splendid!

Well, I don’t have any real guidance to offer, but I can share my glee: for about a year I’ve been hoping DownThem All would be restored to FF or a like product would pop up. My periodic checking yielded nothing until finally, prompted by the above, I just found Download Star.

With respect to SCV though, exactly as I suspected it’s entirely irrelevant as I’m guessing the way audio files are generated involves some cunning magic. :mage:

Interestingly though, I could get it to download the first segment (pli & en) of the first listed sutta:

2 Likes

The magic is all in Bhante’s text segments. A text segment is semantic building block. Upon that simple, magical premise, SC and SCV both rest. There are a lot of text segments in existence and many more to come. In terms of your screeenshot, the technology just implements this:

  1. Every text segment shall have a name. E.g., mn1/en/sujato
  2. Every text segment shall have root text and/or translated text identifiable as such (e.g., “en”, “pli”).
  3. Each language aspect of a text segment shall be speakable and identified with a GUID. These are the mysterious alphanumerics you see in your screenshot. Each text segment aspect has a GUID.
  4. A playlist shall be the concatenation of chosen text segments
  5. Playlists shall be playable or downloadable

The Download Manager might not be practical because what is downloaded are the individual text segment audio aspects. They are quite literally sound bites. What we want is the full course meal presented entire, not spoon-fed. It will take a while to figure out how to apply glue here.

2 Likes

Thanks! Yes, it was never put forward as a solution as I had a vague sense of your above explanation. It was more idle curiosity in view of the knowledge that such a thing as http://50.18.90.151/scv/download/sutta/mn57/en/sujato/recite does exist and work as one whole sutta. Anyway, no need to get side tracked by my idle curiosity. :wink:

2 Likes

This is actually quite interesting and we should discuss what is desired. The word “study” only appears in the “blurb”. Not in the actual translation. Therefore, “study” appears here as Abhidhamma. And to make matters even more interesting, the blurbs are not spoken by SuttaCentral Voice. The blurbs are however available for the blind via screenreaders speaking the first section of any SCV sutta.

Some options are:

  1. do nothing and simply document the issue
  2. restrict searches to text segments only and ignore meta-text about the sutta
  3. play the blurbs in the player (technically a bit of a feat since blurbs are not text segments).
  4. …?

I would be somewhat inclined to choose #2 since it would also address the “sabbamitta” search issue, which is related but not the same. In the case of “sabbamitta” returning four results instead of two, what is happening is that those are the suttas immediately preceding and following the Sabbamitta sutta. Apparently each JSON sutta has links to its neighbors. The links include names, and the name of the sutta in question is “Sabbamitta”. However, in choosing #2, we would take away the ability to search Abidhamma.

1 Like

I’d favour this as well, and it seems quite logic to me: What we want to search for is sutta text, isn’t it? Not the neighbours, not the blurbs…

1 Like

However, if sutta titles were included that would actually be good.

1 Like