zodvik

Nice.

I built https://utils.foo last week. Client side only + ads free. Idea is to keep adding more utils to it, but keep it client side only.

Author here. TIL about the book.

For reference https://www.goodreads.com/book/show/53343.On_Writing_Well

At Stripe, this was also framed as BLUF - bottom line up front.

Please do, that would be awesome.

How does this compare to hurl.dev?

Yep

If you're going to do a single request or a few requests in isolation, curl is the way to do.

If you wanna do a full fledged testing suite, might as well using a proper scripting language or a test framework.

Hurl hits the spot between those two for me.

The abstraction is provide better UX. Anyone will have to write stuff over telnet to provide chained requests, variables, assertions, etc. Once you do that, you basically get a reimplementation of hurl.

Because liking a cli is an individual preference and lots of dev, including me, would prefer it and find it useful.

I saw this comment and thought this exactly described what my $previous_company did. Then, I saw the username :-)

Yes, because most services haven't met the full scope of the RBI circular yet - sending pre-debit notification, self serve mandate revocation, etc

That isn't how it will work

- First charge needs a AFA and it will setup a mandate - Subsequent charges will use that mandate and work without 2FA - Cardholders need to be notified in advance of a debit and be given an option to cancel

What's breaking now for most services is handling the third part

2-random was also used in this HDFS change - https://issues.apache.org/jira/browse/HDFS-8131.

Default block placement policy chose datanodes randomly for new blocks. This change selects 2 datanodes randomly and picks the one with lower used disk space.

Thanks.

This has been my approach thus far. The place where I work allows for practical applications of this (scaling systems to millions of requests per second, having no downtime operations, doing BCP/DR, etc.)

2 things on this - Learning by doing sometimes feels like spending time on discovering things that would have been obvious given the right resources to look at - It's still easier for non-functional aspects of system where there can be quicker feedback. The more abstract part to me is functional design of large systems

For e.g., there are great resources for more "basic" patterns (on Java design patterns, Effective Java, Clean Code, etc.). Are there such resources for roles beyond SSE?

I wrote a lot of code for payments processing at an Indian e-commerce shop & see/make it scale from 100s of QPS to ~200k QPS. Just having the code paths exercised by so many people over a day gives me joy.

One change that I will not forget that made the highest "perceived" impact with a one line change. The payments flow was at a different domain the rest of the site. This meant the users incurred a fresh set of DNS lookup, TCP connection establishment & a SSL handshake cost. The change was to load a "pixel" image from the payments domain a page before. It dropped the first call latency from 550ms to 100ms :-)

FYI - I'm not the author of the article. I found it very useful and hence posted it here on HN.

Hangouts Chat is such a botched launch. It is backwards compatible with single user chat and not group chats. It effectively ensures that people cannot completely migrate to it. The overhead of keeping 2 apps and getting the same notification twice was too much - ended up going back to classic Hangouts.

Would AOF persistence as 'everysec' not imply a write guarantee of max 1 second data loss for restart?

How will infant mortality rise as a result?

You’ll get premium eyewear in India for $150. Though, it will take around 3-5 days to get it after your order. Good quality ones that last would be $60-80.

Thanks to @steanne, here's the link https://news.ycombinator.com/item?id=14497295

Thanks! That's the one.

Tried it on www.flipkart.com (Amazon of India) and it scores 72-75/100. AliExpress scored 78.

Dlang forum (with dynamic content) is insanely fast! https://forum.dlang.org/group/general

Absolutely the same case in India and then some more.

I would argue that Uber (and Ola) have raised my standard of living as far as transport is concerned. I've stayed in Bangalore & Chennai and have travelled in a few more cities.

All problems that you mention ring true here plus an additional things - most taxis & autorickshaws don't even run on the meter. Every ride involves a ritual of haggling the price. If you're new to a city, it is certain that you would be fleeced. Then, many would deny the ride to a destination if it is not a hotspot. On few days I've take the local transport, I get dropped off a kilometre away from my home and have to walk the rest.

I would never take a local taxi or an autorickshaw if there is an Uber/Ola available.

This is neatly done. Thanks to Github for open sourcing this.

Thanks! That clears it up.

What information does the extension send back to your servers?

Yes, long before Amazon started operating in India.

<plug>

We have built a fairly robust system along the idea of Kafka Connect and it's open sourced at https://github.com/flipkart/aesop. Currently, it supports MySQL & Hbase as sources and MySQL, Hbase, ES, Kafka as destination.

If you're using ab or httpperf, you should also take a look at wrk (https://github.com/wg/wrk). It can generate significant amount of load and has support for Lua scripting for request generation, response processing and reporting.

There is another fork of it (https://github.com/giltene/wrk2) which can be used if you want test at a consistent RPS.