anandology

I see the problem. But I feel switching to async is too big a change and there could be simpler ways to address this.

1. Solr was used (I think still continue to be used) for many things it is not quite the right fit. It was done then because, it was simpler to do. For example, it is used to get list of editions when a work is loaded. Limit Solr only to search and figure out a better way to manage the other pieces, would be a good start. Search could be made a different service and the frontend could call it via API. Both of these together will remove Solr from the critical path.

2. The availability status of archive.org could come from a database and a backend task can continue to update it very frequently.

I was one of the initial developers of Open Library and wrote most of it's backend code. I doubt any of the performance issues could be because of code being sync. Open Library is IO bound and it is most likely spending time waiting for db to respond.

Was there any performance analysis done that indicated that the "code being sync" the main bottleneck?

Happy to help.

I have been playing with a Raspberry 400 for couple of days now. I thought of recommend it for my nephew, but couldn't do it because of the following drawbacks.

1. Lack of 3.5mm audio jack or built-in microphone/speakers 2. Unable to run Zoom reliably

Given that schools are closed (at least in India) and classes are happening over zoom, these two drawbacks makes it a no-go for my use case.

> Just for curiosity, is the # of forecasts due to SKU/locations, or due to granularity of series (i.e. let's forecast absolutely every KPI driver, as well as on different time scales). In our case, the big number was due to SKU/locations. But in other use cases, it could be due to multiple KPIs.

> Apologies if that's an low level question, the Walmart blog was relatively large in use case.

I agree. I was trying to highlight the amount of infrastructure and setup required to scale forecasts. While that makes sense for their scale, there are hardly any options to achieve that without dealing with infra.

Didn't realize I could use `itertools.count`. In fact, it accepts a `start` parameter.

The `take` function is borrowed from Haskell. It feels lot more natural to say/read `take(10, primes())`, than `islice(primes(), 10)`.

Any slice implementation that references to the original data has to be some kind of "fat pointer". Is there any other way that can be done?

String.substring in Java does exactly that. I'm sure numpy arrays in Python are implemented like that, perhaps slightly more fatter.

In addition to apt and npm, I would like to see docker image distribution powered by IPFS. It really feels stupid to pull images from a central registry sitting on the other side of the globe when the image is already present in the next node in your kubernetes cluster.

Reminds of Bhopal, a city in India. People there are very fond of numbers. You see people getting on a bus and buying a ticket for "aath number" (number 8) or even "saade paanch" (five and half). You are right, thay are names of places. Not feeling well? get youself checked at "baara sau pachas" (twelve hundred and fifty). Yes, that is the name of a public hospital. It took me a while to get used it.

Agree. Thats what makes jupyter lab in the cloud very interesting.