Fork me on GitHub
#off-topic
<
2017-08-01
>
gleb05:08:27

Sorry for off top guys, but I didnt find any more suitable channel. I need feedback from you but not just as coders rather as product managers and product creators. In the short post below I’m talking about 10 exercises to pump your product thinking like biceps and get the brand new ideas for the projects you lead. Please check this out. 💪 JUST DO IT! — https://medium.com/standuply/10-exercises-to-train-product-thinking-cda9882327b7

Sy Borg06:08:36

ok, my feedback - I don't like this bullshit here.

tonic440110:08:07

Agree with sy_borg

val_waeselynck13:08:59

@U57THKJ1X I would say those seem more like principles than exercises 🙂 IMHO it would be interesting to describe how things go wrong when you don't follow them. I'm also personally skeptical about phrases like 'treat life as a project' - IMO it's a rather bad philosophical habit that we see a lot in the tech / startup world to generalize some work methodologies to rule our lives.

qqq12:08:53

I'm looking for something that can convert 3.5 hotswap drive to 2.5 hotswap. On amazon, https://www.amazon.com/s/ref=nb_sb_noss?url=search-alias%3Daps&amp;field-keywords=2.5+hotswap+3.5&amp;rh=i%3Aaps%2Ck%3A2.5+hotswap+3.5 all have 'extra electronics' where the 2.5 plugs into, and then it fakes the 3.5 output my question: why can't the caddy just place the 2.5 in such a way so that the sata from the 2.5 goes directly into hotswap ?

alandipert18:08:58

speaking for myself, as someone who helped with the production of brave and true, i can appreciate and respect that some people think its stupid. never heard of a tech book without detractors

qqq18:08:22

SICP 🙂

alandipert18:08:30

i guess i would prefer to spend my oss time communicating with other people whose criticism is more helpful than it is rude

alandipert18:08:49

but this being off-topic, i'd say we're all collaborating excellently!

qqq19:08:47

@alandipert : it sounds like I missed the thread; just to be clear, I'm not defending whoever that was criticizing 'brave and true' // I was merely asking "tech book without detractors" -> SICP

dpsutton19:08:10

well for Brave, big big ups for it being free on the internet. I bought a copy last year so i could read it on a plane. I think its excellent also because it really focuses on the pure side of the language

dpsutton19:08:27

i'm thinking of the way the game was developed with triangular numbers to recover state

dpsutton19:08:07

and writing humorous and engaging technical prose is hard. And it finds a really good balance of humor to content

mobileink19:08:53

harder than coding imho!

alandipert19:08:30

qqq hehe no worries :-) i just chose not to respond because i have major beef with sicp. alice p. hacker is a preposterous name

dpsutton19:08:28

who's the other one? Eval Uator?

dpsutton19:08:30

something like that?

dpsutton19:08:40

always noticing the problems with the faulty persona

dpsutton19:08:47

Eva Luator?

alandipert19:08:58

dpsutton thanks for the compliments, i cant claim to have written. it was all daniel, i just fixed some typos. but ill pass it along! agree that humor/content = very hard

qqq19:08:25

wait, what are the technical complaints aganist sicp ?

dpsutton19:08:40

> alice p. hacker is a preposterous name

alandipert19:08:08

in seriousness, i am nursing Doubts that lisp is a good thing to teach people first cracks open can of worms, passes around

dpsutton19:08:28

also that you build up quite some big systems and they leave lots of things undefined until later. its often not clear when you have all of the code required

dpsutton19:08:33

like the constraint system

mobileink19:08:13

Best is https://mitpress.mit.edu/books/little-schemer. If you can stomach the godawful cuteness.

dpsutton19:08:22

ha. my coworker is looking at brave online. I'm leaving and there's a production system in clojure here and they need to learn it

qqq19:08:22

@dpsutton : I thought that was by design, the notion of "wishful thinking" ; if we only had a function for FOO, we can build BAR; okay, let's write BAR pretending we have FOO, then we will write FOO later

dpsutton19:08:39

it is. but its easy to get lost in it as well

dpsutton19:08:54

and they will give full fledged examples of the product before its fully defined

dpsutton19:08:05

just a small critique about usability. love the book still

tbaldridge19:08:06

Its also super abstract. SICP dives into things like prolog unification and meta interpreters without touching on things that are much more useful first.

tbaldridge19:08:33

You can know everything in SICP and still be utterly clueless at writing production apps. And that's kindof my problem with it as a starter text. Sure I love some of the content, but that's because I enjoy working on compilers and interpreters. For the average developer something a lot more concrete would be better.

tbaldridge19:08:04

Yes a prolog interpreter is for the most part walking two trees at a time, but maybe we should start with thoughts about why we would need to walk two trees at once (hint: we rarely do).

qqq19:08:35

iirc, unification doesn't happen until chapter 4 (out of a 5 chapter book)

alandipert19:08:59

re: gearing up to write production apps, that seems like a set of topics that would succeed the kind of theoretical topics sicp was designed to illuminate

qqq20:08:07

what is this bitcoin cash split thing that people are talking about? on http://coinmarketcap.com I only see one bitcoin ticker

noisesmith20:08:53

it’s a split of the network, some people want more room for data so they are forking the blockchain and code - “bitcoin cash” is that fork http://www.businessinsider.com/bitcoin-price-fork-happens-2017-8

mobileink22:08:48

do you feel lucky, @qqq?

qqq22:08:07

I don't own any bitcoin.

mobileink22:08:00

I just bought $150 worth. Just for fun. Except, it is not clear the transaction went thru. Some of the BC places put a halt on transactions until the fork kerfuffle blows over.

qqq22:08:46

oh wait you bought $150 bitcoin cash (the new one) ?

qqq22:08:34

since there is a fork, are the people who owned bitcoin before the fork able to spend on both branches ?

mobileink22:08:41

no, not the new stuff. but I think I’ll give it a try. I can afford to lose $150.

mobileink22:08:44

i think it depends on whom you do biz with. some places are not supporting bitcoincash. frankly, it’s all over my head.

mobileink22:08:34

all I know is if I had spent $1000 on BC a year ago I would have made 350%.

mobileink22:08:17

McAfee thinks one BC will be worth $500,000 in 3 years time, and offered, I am not making this up, to eat his own dick if that does not happen.

mobileink22:08:56

he might be crazy, but what if he’s right? that’s worth a $150 investment. 😉

roberto22:08:37

you are going down a dangerous path if you start to give McAfee an ounce of credit on any of his comments.

mobileink22:08:06

but he’s rich, I’m not. he must be smarter than me, hah hah.

roberto22:08:27

you might as well ask the mooch for advice then 😄

roberto22:08:55

isn’t McAfee bankrupt too?

mobileink22:08:05

hey, the mooch may be a douche-bag, but he also made himself very wealthy. they do not let idiots into Harvard Law.

roberto22:08:13

hahahahaha

roberto22:08:30

Harvard lets anyone in that can pay up

mobileink22:08:47

Erm, I don’t think so.

roberto22:08:50

didn’t Kushner also go to Harvard?

roberto22:08:31

and apparently he was not bright at all based on accounts from his contemporaries at Harvard

mobileink22:08:34

So did Bush the Younger. (But that was Biz School, not the same.)

roberto22:08:00

what makes Law School the exception?

roberto22:08:09

because Lawyers have ethics?

mobileink22:08:17

ok. but “not so bright” among Harvard Law does not mean “kinda stupid”. It means more like “not exceptionally brilliant”.

roberto22:08:28

I was trying to be polite

mobileink22:08:15

Everybody in the big leagues sneers at the biz schools.

mobileink22:08:24

Pretty sure Kushner is very very intelligent. Ditto for the Mooch.

mobileink22:08:35

Wise is different.

mobileink22:08:36

anyway, topic is bitcoin. fascinating to watch.

qqq22:08:28

topic is off-topic; it's even okay to discuss things like polymer/miraj

mobileink22:08:56

wonder why that never seems to happen.

qqq22:08:55

I bought an poweredge server; it's cljs build time is slower than my i7

qqq22:08:06

a refurbished poweredge to be exact

qqq22:08:18

I wonder if it's 1. i7 beats dual xeon or 2. cljs build is highly single threaded

mobileink22:08:57

did you flip the cljs toggle?

mobileink22:08:21

is it also slower on other tasks?

mobileink22:08:19

e.g. building cljs master?

qqq22:08:27

you mean :parallel-build true ? yeah, that went from 70s on poweredge to 60s on poweredge; but a $500 desktop machine w a 3.4 ghx i7 can compile in 50s

qqq22:08:42

I haven't tested other tasks; I'm just using whatever cljs boot is using

mobileink22:08:07

well poweredge covers a lot of ground

mobileink22:08:30

i’d test some basic java tasks

mobileink22:08:51

a cheap poweredge gets you a Celeron cpu, no?

qqq22:08:16

no, they're zeons

qqq22:08:41

they're refurbished was-top-of-the-line 5-yrs-ago rack server machines

mobileink22:08:45

hmm, I’d like to know the explanation. lots of “xeon v. i7” articles on the google but they’re over my head.

mobileink22:08:44

there are so many variables involved

mobileink22:08:33

i.e. there’s more to it than just the cpu.

qqq23:08:11

I suspect the poweredge is optimized for throughout whereas the desktop is optimized for latency

qqq23:08:28

and thus, the desktop is beating out the refurbished polweredge for cljs compilation

qqq23:08:42

if I were to benchmark postgresql, it'd probably be different results

mobileink23:08:44

yeah, something like that. from googling it looks like the cpus are pretty similar.

mobileink23:08:39

i’m thinking memory stuff, caching, etc.

qqq23:08:58

ah right, the memory is only 800mhz

qqq23:08:07

that may be a contributing factor

mobileink23:08:41

some interesting articles if you google “why does i7 outperform xeon when compiling”

qqq23:08:19

was just reading precisely that

mobileink23:08:36

leaves me just as confused, frankly. 😉

qqq23:08:49

yeah, I'm not a big fan of thos3eanswers

qqq23:08:06

what I got so far is: xeon supports ecc memory i7 has in proc video card

qqq23:08:13

xeon better for VMs too

mobileink23:08:25

you’re using boot, yeah? so it’s going to depend on jvm performance either way. maybe the jvm doesn’t know how to use the cpu, so to speak. just guessin

mobileink23:08:59

have you tried alternative jvms?

qqq23:08:34

great question

qqq23:08:39

what jvm am I using

mobileink23:08:54

btw, just how much slower is it?

qqq23:08:52

core i7 3.4 ghz + 32 GB: 40 seconds poweredge 2x6 core 2.0 ghz + 96 GB: 60 seconds poweredge 2x6 core 2.0 ghz + 96Gb + parallel-build: 50 seconds

qqq23:08:21

java -version openjdk version "1.8.0_131" OpenJDK Runtime Environment (build 1.8.0_131-b12) OpenJDK 64-Bit Server VM (build 25.131-b12, mixed mode) (fedora 26)

qqq23:08:28

maybe I need to install sun oracle java and see what happens then

qqq23:08:37

@mobileink which one are you suggesting I use ?

mobileink23:08:13

that page says oracle and openjdk use the same jvm - HotSpot. Dunno if that is true.

mobileink23:08:49

no idea. not even sure that’s the problem, just a possibility.

qqq23:08:52

55s openjdk // 53s sun-oracle

qqq23:08:42

I'm going to stick to the theory of poweredge-xeon = optimized for throughout; desktop-i7 = optimized for latency

mobileink23:08:43

what about your hard drives? reading stuff from a HD is slow anyway, if your i7 has much faster HD access that could explain some of the difference.

qqq23:08:59

they're both using some very fast ssd

qqq23:08:18

poweredge is even raid0, which I believe uses both ssds, which should reduce ssd latency in half

mobileink23:08:42

hmm, well, I’ll be interested to hear what you figure out. I’m flummoxed.

qqq23:08:08

typical @mobileink , increases # of questions without providing more answers 🙂

mobileink23:08:31

Snatch the pebble from my hand, grasshopper. 😉

qqq23:08:32

how's your work coming along?

mobileink23:08:03

slooooowly. ups and downs..

mobileink23:08:54

you can only code in white-hot mode for so long, then you burn out.

qqq23:08:40

yeah; I'm starting to 'code' on paper/pencil alot these days, and sit at a keyboard only when I have to type things out

mobileink23:08:17

i’m working out the transition from polymer v1 to v2 (proper web components, technically v1). but only in my head. sometimes you just cannot write code, even when you know what code you need to write.

mobileink23:08:25

knuth wrote TeX with paper an pencil. Typed it in, and it worked. ha!

qqq23:08:14

Michangelo once took a block of marble, removed some parts of it, and poof, out came michangelo.

mobileink23:08:34

wow, recursive sculpture!

qqq23:08:27

hmm, it may have been david instead

mobileink23:08:29

that should be an O Henry short story.

mobileink23:08:50

no, better if it is Michealangelo!

mobileink23:08:18

problem with recursive computation: how does it end? problem with recursive sculpture: how does it start?

qqq23:08:57

I may have rediscovered a law of nature: conservation of compile time: regardless of how much faster computer gets, compile times stay the same: because when it hits 5-10s, it's "fast enough" and devs stop optimizing the compiler

noisesmith23:08:39

see also battery time in mobile devices

mobileink23:08:46

or maybe a constant.

noisesmith23:08:57

mobile battery life will always be “almost enough for one day, a full day if you are careful”

mobileink23:08:12

and never enough when you need it!

noisesmith23:08:19

when the batteries get more powerful, the devs get sloppier and users start doing more inherently power hungry tasks

mobileink23:08:52

or: don’t worry about it, the hardware will get faster

qqq23:08:07

@noisesmith : I wonder if there is a market for a minimalist laptop that just ran ssh and had a month of battery life per charge

noisesmith23:08:55

I for one think that would be pretty useful

mobileink23:08:05

chromebooks?

noisesmith23:08:07

the reinvention of the thin client tty

mobileink23:08:37

there’s a mkt for that. I’d buy one. I’ll offer $200.

qqq23:08:01

how power hungry are 5g chips ?

noisesmith23:08:48

with a thing like you are suggesting, the biggest power issue would likely be the wifi / 4g (if it supported either) - the other stuff could be brought down to very low needs

qqq23:08:18

probably, especially if other devices are blasting wireless signals as if they only ayve to last for a f3ew hours

mobileink23:08:02

add usb 3.0, I only need a reliable full day of battery

qqq23:08:05

we need bio computers that output directly to retina, takes input from brain, and runs on human fat

qqq23:08:20

not sure how it would be wifi though

mobileink23:08:45

it’s called Energy Harvesting and it is a very active research area: https://en.wikipedia.org/wiki/Energy_harvesting

mobileink23:08:36

before too long the only thing you will need to do to charge your phone is walk.