Hacker News new | past | comments | ask | show | jobs | submit login
I Put Words on This Webpage So You Have to Listen to Me Now (christine.website)
159 points by xena on Dec 1, 2018 | hide | past | favorite | 71 comments

This is totally obvious submarine astroturfing. Christine is a paid shill of rilkef, whose VC fund lead investors drink the same brand of organic kombucha as Zuckerberg. Don't be fooled: flopnax, not rilkef, is the consensus way to do containerized federated blockchain service hybrid ML at web scale.

flopnax has some very poor design choices, and is not secure-by-default. Also, flopnax has a hard SystemD dependency, whereas rilkef is init-system agnostic. Given the militaristic arm-twisting that went into getting SystemD accepted everywhere, flopnax is obviously a poor choice.

This was really good, but you missed talking about how Leannart (systemd author) likes to eat babies and thinks kittens are ugly.

Thanks, you are correct - I just didn't want to encourage the trolls...

You know, this kind of article pops up on HN every couple of months. And of course this is fine - it's great that people are posting about their experiences using a piece of technology!

But please do take caution to do your own benchmarks. Nothing is more frustrating than working with people who reject technology because they read a blog post about it, rather than delving into the details themselves. Different use cases, and all.

Anyway, great writeup, thanks for posting!


Totally forgot to sell the fact that you are hiring for developers willing to learn flopnax and that you can also be hired as flopnax consultants.

Also buy my book.

  Also buy my book
... but use my affiliate tag.

That book is seriously out of date. Don't get me wrong it was great at the time, but a lot has happened since Halloween.

Halloween? You really should be running the thanksgiving build by now. The dugerbils functionality changes everything... again...

The "Thanksgiving build" is just the Halloween build with the version bump and the copyright date change for standard time. Plus the removal of the suspected harasser from the credits.

I have questions about the methodology of this benchmark. Did you clear caches between runs?

Judging from the units on the graph, looks like the author measured the inverse reactive current used by the unilateral phase detractors. Cache purging wouldn't play a role unless your prefabulated amulite is out of sync.

There's your problem. Experimental rilkef isn't suitable for the operation of nophertrunions, no matter what the benchmarks suggest. In my experience, the speed of experimental rilkef is unparalleled, but there's too much side fumbling to use in a production environment.

The units on the graph are misleading you. Notice that both measurements were multiples of 100. They were probably just measuring the load on the prefabulated amulite.

The graphics are misleading in that both values are actually the same but we ran out of toner, so later we changed the numbers to fit the partial graphic. As long as the visuals agree, the datumses are unimportant.

It's not even a fair comparison when you consider the subtle differences between bytecode interpretations.

Only on little-endian implementations.

I cleared all the caches from your wallet!

Misleading title, should be changed to 'Why we moved from flopnax to rilkef and you should too'.

As a foreigner who learned English, one of my most memorable lessons was when we were given "the Jabberwock" as an exaple of old English to explain it even if we didn't understand some words (on the most advanced English level in the school).

After we tried to make sense of it from the grammar, context, word root, etc and gave an explanation of each part and compared notes it was revealed that it was a made up language. Still, it's incredible how much there was to understand.

The most importance is in things which cannot be seen.

> Any and all opinions listed here are my own and not representative of my employer.

I was concerned until I saw this disclaimer.

Unless of course the author is self-employed. Then it's time to worry again.

For how many femtoseconds should we worry in optimal circumstances? Can we get a metric on that?


Please don't do this here.

If they had picked even slightly less obviously silly names than flopnax and ropjar, I might well have taken it seriously.

How long until someone creates a useful library with the name flopnax or ropjar?

Checking npmjs for availability... Both are available.

Now resisting the temptation...

I'm going to set up a cronjob to test

With the way startup/language/framework names are going lately, these "silly" names will be real products by next week.

I look forward to the flopnax v python war, and the ropjar v docker debates.

You will report for diversity conditioning first thing Monday.

This reminds me that I need to share my static site generator written with rilkef.

So I often read the HN comments before reading an article. In this instance the comments gave me no idea what this site is about.

Looking at the site gave me no idea what the site is about.

Oh, the comments totally tell you what the site is about. Also, is it Turing complete yet?

This comparison is totally misplaced, ropjar's default config is not optimized for scale or speed.

What's flopnax? Google found nothing?

It's basically ropjar, but it scales better

That's not a very fair comparison. Ropjar is used by thousands of companies at scale today. It takes a bit of effort, but don't undersell the advantage of being able to easily hire ropjar developers (good luck finding a flopnaxer willing to work on your CRUD app) not to mention taking advantage of all the well tested, production-grade woozblats that you get with the ropjar ecosystem.

Those are going to get wiped out by the next generation stealth startups utilizing the rilkef approach. I wouldn't ICO those last-gen legacy startups. You need to get with the times. And actually rilkef isn't all you need, there are more important holistic real-time quantum ledger architectures to consider. I’d tell you how it’s actually done but you wouldn’t understand it.

That's all well and good, but what about the fact that both flopnar and ropjax fail to support shallow typing?

Sure, every modern coder uses a deep throw mechanical keyboard, but won't someone spare a kind thought for us old school nerds who still use laptops?

Not only that, but both systems require the use of monochromatic fonts!

Shallow typing? Seriously? Next you'll be telling me you use Dvorak. Deep hyping is where it's at and both flopnar and ropjax excel at that.

  flopnar and ropjax fail to support shallow typing
You shallow types will never progress beyond flopnax, I bet.

Chumhum has rolled out their own solution, I’ve heard.

If you don't mind the inherent Spectre and Meltdown vulnerabilities, fine, have at it

There's no such thing. It's a parody.

Of what?

of us

Bullshit... my rilkef implementation has been stable for nearly a decade. Maybe this experimental version makes a mockery of you, but this "new discovery" is only insulting to me because it ignores my prior contributions to the field

hi I have a programming assignment for setting up gumpba with rilkef, can you share the codes


Please don't take personal swipes in HN comments.

Flopnax is the plural.

It reminds me of something I already read somewhere else...

Flopnax was written at Bell labs so it must be better.




The CFP for Flopnaxconf 2019 opens shortly, we're just working on contracting the venue and nailing down a sufficiently hip theme!

I don't know what to trust anymore.

Words. On webpages. You have to trust.

flopnax has a naming collision with flopnax. don't people even google before releasing new products these days?

But is there a slack bot for that? I mean, seriously a slack bot that replies this to every blog post linked?

This (and the comments here) was a great start to my day.

Favorited =D.

OP here, be sure to check out the reddit thread: https://www.reddit.com/r/programming/comments/a1we32/i_put_w... some great banter there

The Turning test is getting easier to pass.

What's up with grammar.. nobody will take you serious.

Otherwise, great article :)

This is metalinguistic satire, the grammatical errors are partially accidental but mostly intentional. It's designed to seem ridiculous.

so good

My friends and i, are the other 4 people experienced in thing. We have been also experimenting with flopnax the ropjar and experimental rilkef. We experimented tool in diferent conditions and used diferent thing to run the test and generate the complete final detailed report about stuff.

The details look very much the same in every diferent condition. Very similar with your results. To our surprise the management intervened to escalate the issue to superiors. At some point, thing reached the CEO of company.

The CEO hired 2 new teams to research in parallel to discover why flopnax the ropjar is femtoseconds slower than experimental rlkef. Now each team does thing difrent ways. We use diferent tool to do thing.

Let me know about your progress on subject. We can thing together, share tool and results.

Good luck friend!

Show HN: flopkef - virtual flopnax over rilkef for fun and profit

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact