thanatos_dem

What’s going on with that first graph? The scale changes from 1500” to 3000”, but Spark’s bar sizes stay the same.

Did Spark have a 2x slow down between the tests? It seems pretty misleading.

Because crapping on Mongo is “in” right now. Years ago it used to be “in” to crap on it for its deaign choices, especially its tendency to lose data (seems to be what the parent comment is referencing; see [1] for a summary of some issues from 2013).

More recently it’s “in” for their business decisions, with the relicensing and the rise of API-compatible alternatives pushing some users away.

But yeah, really no reason to mock them in this discussion, apart from going for internet cool points from an audience that may also be critical of Mongo.

[1] https://aphyr.com/posts/284-jepsen-mongodb

...Are you really trying to use “gophers” as a derogatory term?

I think the parent poster’s point is that that’s a very western perspective of morality. Imagine if the tables were turned, and the apprent promiscuity promoted by apps like Tinder was deemed immoral by a country that’s more conservative. Should that app be pulled globally as a result?

To them the fact that we allow so much casually dating could be (and in middle eastern cultures often is) viewed as “wrong”.

> Just look at how the opinion about stretching changed

What happened with stretching? Did we formerly not like stretching, or did we formerly like it but no longer?

Reading through this, I kept thinking that ULIDs[1] give the same benefits described, with wider adoption/support.

Luckily it looks like the author has already written up his thoughts on the differences[2].

[1] https://github.com/ulid/spec [2] https://github.com/segmentio/ksuid/issues/8

> Wow, what a condescending view point that people can't possibly think through things on their own without other people making decisions for them.

...

> Smoking, seatbelts, and privacy are all personal choices that people should be able to make for themselves.

There is a legal minimum age for purchasing cigarettes. Most states have laws requiring the use of seatbelts. So why shouldn’t some power, be it governmental or corporate, push people towards protecting their privacy?

If you’re going to cherrypick counterexamples to the idea that people can think things through and make the right decision for themselves maybe try picking, ya know, counterexamples?

On mobile it’s also responding to the accelerometer data. At least on my iPhone X. Very cool demo.

“Infinite horizontal scalability”. Yeah, no. There’s coordination overhead for adding more nodes. It needs to have an upper bound. It may be a high upper bound, but infinite is straight up impossible.

It may seem trivial, but claims like that are why I prefer CockroachDB when it comes to NewSQL systems. They’re at least very upfront with its compromises and areas needing improvement.

I’ve never see anything from Pingcap explaining the compromises that the DB makes. A “when should you not use this” type overview, and to me those are the most important things when deciding on a technology.

It may be good technology, but it’s marketed like modern day snake oil.

These look awesome, but I cannot imagine spending $1,400 on a clock...

Probably to keep the clock sync'd?

Nope, that one is all you

Some design decisions aren’t great. For instance, given a service key (effectively the identifier for the service to page), there is no way to tell if it’s for use in the v1 or v2 API. No identifiers in the key that you can check offline, and nothing in the API that you can check it against. They are just uuid’s with the hyphens removes, which is a sort of bizarre formate. So if you have a service that generates pages with user provided service keys (i.e. any custom integration), you either need users to also specify which API version it’s for, which is prone to errors, or blindly try one API version and then if it fails try the other, which increases complexity.

Not a deal breaker by any means, but indicates to me a relatively inexperienced team designing the APIs.

Pourque no los dos? Foundation announced a mongo api compatible document layer in November - https://www.foundationdb.org/blog/announcing-document-layer/

It’s a natural byproduct of always looking for the next big thing, no? You need to take chances and try new things, and not all of them work out.

The intro to low level computing class at Georgia Tech used Game Boy Advanced as the platform that the projects were built on.

It was cool at first, but I ended up too deep in the weeds with platform idiosyncrasies to really get much actual C knowledge.

Yeah, along with the FDB document layer: https://www.foundationdb.org/blog/announcing-document-layer/

I know it’s an exaggeration, but it’s also a (small) possibility. Believe it or not I’m not at my most rational in the middle of a panic attack ‍️

Anecdotally, marijuana has been an amazing improvement to my mental health.

I have suffered from insomnia my entire life. Then two years ago, my job got significantly more stressful and I began having panic attacks. I was perscribed Lorazepam in case of future attacks, but the potential side effects were terrifying (you may just suddenly die in certain circumstances), so I was scared to take it.

Enter medicinal marijuana. I’m now part of the “smoke weed every day” lifestyle, except with edibles because my apartment doesn’t allow smoking. 20mg CBD, 5mg THC, about 2 hours before I want to sleep. Accept that I’m not going out for the rest of the night, work on some side projects, watch some TV, and then go to sleep waking up refreshed, no hangover or grogginess.

Only real detrimental effect I’ve observed is a reduction in my short term memory. So I started keeping a proper notebook and calendar to compensate for that. After a couple of days of not consuming, my memory seems to return to normal, so I’m not too worried about long term effects there, but I would absolutely love to see more clinical trials around it.

Interestingly, SQLite gets it wrong, returning 2.0, but MySQL, MariaDB, Postgres, and Cockroach all get it right at 1.0

I guess this comes down to most of them having implementations of arbitrary precision decimals.

I work on CI/CD systems, and that’s one thing that definitely gets harder in a monorepo.

So you made a commit. What artifacts change as a result? What do you need to rebuild, retest, and redeploy? It doesn’t take a large amount of scale to make rebuilding and retesting everything impossible. In a poly repo world, the repository is generally the unit of building and deployment. In monorepo it gets more messy.

For instance, one perceived benefit of a monorepo is it removes the need for explicit versioning between libraries and the code that uses them, since they’re all versioned together.

But now, if someone changes the library, you need to have a way to find all of its usages, and retest those to make sure the changed didn’t break their use. So there’s a dependency tree of components somewhere that needs to be established, but now it’s not explicit, and no one is given the option to pin to a particular version if they can’t/won’t update. This is the world of google & influcenced the (lack of) dependency management in go.

You could very well publish everything independently, using semver, and put build descriptors inside each project subdirectory, but then, congratulations, you just invented the polyrepo, or an approximation thereof.

Unless I’m missing something, your second example seems to be the one in the article.

Did the author change it?

Edit: just saw the comments on the article. Yes, she did.

I think the point of the “constant” is that the average hasn’t changed over the course of time. So it’s a constant average. Not like a ‘const’ variable.

The company I was at at the time was very startup-y. Fast paced, high stress. People in the office very late, and gave a stipend for living walking distance to the office.

The result was a perception by myself and many others that we needed to stay later, and could because hey, home is right around the corner.

I’m sure some people could manage that better, but a lot of the company, myself included, could not.

When it was South Bay to SF I commuted by driving to Caltrain, taking the train up, and then walking across half of SF.

Now I bike most days, unless the weather is crummy or the state is on fire again.

The work life balance issue was definitely partially the fault of the culture of the company I was at at the time, but living so close made it much more difficult to avoid.

And yeah, that’s an issue of willpower in part, and is the same reason I am not able to work remote or from home. But I don’t think that’s an issue unique to me.

Having had a 3 hour round trip commute time (South Bay to SF), and a 10 minute round trip commute time (3 blocks away), I think that my current time of about 1 hour is the most desirable.

The long commute took a toll for obvious reasons, but even with the short convenient walking commute, I found it led to difficulties maintaining a healthy work/life balance.

An hour total commute (30 minutes each way) is generally enough to keep work life separated from personal life, while not being too much burden, in my opinion.

Not necessarily, depends on what it’s a screenshot of. If it’s of a web page or application, odds are that it will have large sections of uniform color (white backgrounds for instance). This is compressed very well in PNG, which uses DEFLATE/LZ77.

So it’s pretty likely that the PNG will be smaller than the converted JPEG in many cases. Screenshots of videos, photos, however will likely be converted.

> Each entry contains a bunch of IPFS address pointers - to the last entry on that "database", to any extra entries you found out about, and to a bunch of previous entries (just to speed up reading)

How bad is the latency to read the entire database state? Even git can become pretty slow fetching and processing objects on large repos, and all of those refs are local after the initial fetch.

Seems like with the additional latency on IPFS, resolving the entire DB state (for snapshotting it or otherwise backing it up, for instance) would be unusably slow for anything even somewhat large, no?

What an abysmal ad experience. Tried to load the page 3 times. Every time had an inescapable pop up ads come up, redirecting me to a spam site, and then redirecting to the same site as many times as is allowed before the browser intervenes, so hitting back does nothing.

This is why the world is collectively joining the add blocker train. Talk about an industry that can’t die soon enough...

> Hint - it can and did end up blowing up in spectacular way.

Very true. See George Soros and the Bank of England for a famous example.