Gemini search

Created 2022-09-19

On 2022-09-14, Metamatters wrote:

It can be tricky to find the good stuff in Geminispace. ... There are many capsules with interesting static content not dribbled out in regular updates, as well as a variety of interactive services, mirrors, databases, and so forth. Search engines help, but you have to know what you're looking for.

thegonz: Announcing the Collaborative Directory of Geminispace

In addition to that, I took a look at a search engine:

geminispace: Gemini Search Engine

and tapped in the search for "stm32" (a small mirocontroller). The search results are at:

geminispace: search results for stm32

There are 38616 results, so I'm not going to be able to search through all of them. There's a few problems. I get the link, but not the title. The text shown in the results didn't seem especially helpful, either.

So OK, suppose I'm interested in how to configure the PWM (Pulse-Width Modulation) hardware that's on the stm32. I search for "stm32 and pwm", but I still don't get a useful result. One of the top links is to "Open Hardware: Arduino, Raspberry Pi, and ESP32", none of which is applicable to the stm32.

Moreover, I'm never going to track through tens of thousands of links. Something more manageable is required.

Returning to CDG (Collaborative Directory of Geminispace), I'd say I'm a bit "meh" on their classification system.

We then have the problem of taxonomies. Fortunately, there is a system that has "solved" this: the Dewey Decimal System.

Youtube: Dewey Decimal Classification Session 1 - June 2010

The DDC (Dewey Decimal Classification) is arranged my discipline, and further subcategorised into subject. An example the video gives is about dogs. How would you classify a book about dogs? Well, the answer depends according to discipline. If you mean in the zoological sense, then it goes into section 591. If you mean in the sense of caring for pets, then it should go in section 636, which is for animal husbandry.

So perhaps we should think in those terms. From the video, it is clear that there is extensive documentation on how to classify book contents, what do do in case of multiple subjects, and so on. The problem? The broad categories are easy to get hold of. The nitty-gritty details - the ones that we'd actually need to drill down into to be of use - are difficult to get hold of. This would be a great resource to have online. Needless to say, though, it seems nigh on impossible to find this information, assuming it was available for free anyway.

Ah, the good ol' internet, an information system that spans the width and length of an ocean, but is only an inch deep. The more Javascript we throw at it, the less like a reference work it seems, and more like a never-ending parade of TV commercials.

So maybe the big idea for Geminispace would be that we would think of it as a combination of blogs and books. A capsule would be organised as a collection of books, which you could tag with dewey numbers that could be picked up by a search engine. Each gemtext file could be like a chapter or page in a book, or even an entire book in its own right. I would discourage the notion of a monstrous page, though.

Just some food for thought. I don't really have much in the way of proper answers.

Update 1

I've been thinking about it, and I don't think the DDC is a very good one.

For example, 005.133 is General Programming Languages, so you'll find Perl in there. 005.135 is Assembly Languages, which seems reasonable.

But then suppose we want to learn about programming various microcontrollers. We have:

librarything: example result

Wow. OK, so 005.13 is close to 005.133 I guess, even though they are both about MicroPython and the Pi Pico. But how did we get to 005.262 for the STM32? The topic is the same, but it's just for a different board.

Then, of course there's 621.38... . I didn't make that long number up. It's actually real. How comes an Arduino cookbook is in a completely different section, so far away from other books in which it seems related? 621 is for applied phsyics. More specifically, 621.3 is for Electrical, magnetic ... COMPUTER ENGINEERING ... .

The DDC does suggest details of how to resolve conflicts and sucklike. The problem is, though, that this doesn't seem to be readily available to the public without a paid subscription. So this raises the obvious question: how can I be expected to use a classification system when I don't even know what the classification system is?

So, on reflection, I think the CDG is a good idea. What it will likely need in future is caveats and revisions, and explanatory notes. We should probably not need to worry about things at this stage, but just anticipate that it will be an emergent property if it really takes off.

I might be inclined to suggest that a gemfile be only once in the index. If its placement is ambiguous, then we may need to introduce explanatory notes as to how contents should be ideally positioned.

I guess there's still a lot of difficult-to-resolve issues. Like, take the subject of Integrated Circuits. It's hardware that may have rarely anything to do with microcontrollers. So presumably they could be in a completely different section to programming. Yet some ICs are very much like microcontrollers in that they really belong in the realm of programming, having things like SPI interfaces. So you might consider them to be software-like items. Or maybe you think microcontrollers belong in a separate category from software as they are hardware with IC-like capabilities, and regard it only as incidental that you have to use programming languages to get them to work.

Ah, the problems with taxonomies.