Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
meaty
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
77 ms
·
751.
▲
by
meaty
14y ago
I do this. In fact, for a lot of what I do, assertions are good enough to catch any issues early. That and integration tests are probably more valuable than religious unit tests. The advantage with assertions as well, is if you leave them o
752.
▲
by
meaty
14y ago
If you need lots of mocks, it's a sign of coupling which means a bad fundamental design. You need to test both the components individually (unit tests) and their interaction (integration tests). We use Commons VFS for filesystem abstraction
753.
▲
by
meaty
14y ago
We do indeed explain that to them and provide them with material so that they can improve rather than just kicking them out of the door. We're good like that. We just can't hire them. Most programming jobs are very involved. Most vocal po
754.
▲
by
meaty
14y ago
No that's not what I suggested. Read it again a couple of times, carefully. I think it was a logical and concise argument. Perhaps a better statement is: "any semi technical skill takes way more than 10 weeks to become productive and not a
755.
▲
by
meaty
14y ago
I get what you're saying (I designed a 4-bit CPU at university, prototyped it in basic TTL and got it fabbed and it worked (!) in two semesters) but it doesn't mean I'll ever be able to throw a useful microcontroller core out of the door or
756.
▲
by
meaty
14y ago
Yes like NEETS as a concrete example which is very very good: http://jacquesricher.com/NEETS/
757.
▲
by
meaty
14y ago
The military, well the British Army which is my only experience (not strictly true [1]), does suck at training. I mean really suck. Most knowledge is pumped down via training manuals written by people detached from the equipment or scenario
758.
▲
by
meaty
14y ago
In military boot camps, you learn enough not to get killed or court martial, not how to do a job properly. (ex British territorial army) I'm not saying it's not possible, but the marketed result versus the actual result here is disparate.
759.
▲
by
meaty
14y ago
In my case, a simplified view: reading the documentation, building a list of scenarios based on proposed functions and associated risks, testing each one and documenting it in a concise manor. Then building something serious, testing it, br
760.
▲
by
meaty
14y ago
No, I believe learning a skill is an investment, which is why I take my time (sometimes years) and study for a long time before I even put pen to paper. My latest learning exercise was postgresql. I've done a proper engineering evaluation o
761.
▲
by
meaty
14y ago
10 weeks is not enough time to learn any skill. I wish people would stop these boot camps with ridiculous claims. We've had a couple of victims of these schemes apply to us and we've had to rather sadly tell them 'no way' and explain why.
762.
▲
by
meaty
14y ago
I'm actually surprised any DC would take that equipment. They, in my experience at least, are very fussy about what what you put in the racks and power draw etc. Oh and we get 640 cores in 20U (8x4 core xeon machines each 1u) and that leave
763.
▲
by
meaty
14y ago
+1 for this. Go can hammer out requests spectacularly. We're using it to shift out static content (all from RAM).
764.
▲
by
meaty
14y ago
Excellent. I look forward to reading about that :)
765.
▲
by
meaty
14y ago
Scan rate is shit though if you have to use spi/i2c.
766.
▲
by
meaty
14y ago
Nothing wrong with facts. I'd rather get a balanced view of someone than the spin. Look at what they did to Steve jobs! The guy was an asshole. I dont celebrate the life of assholes.
767.
▲
by
meaty
14y ago
The PCB is the easy bit. Making the firmware which delivers your keypresses down a PS2/USB pipe is not. You have to consider things like n-key rollover and ghosting/blocking particularly when using matrix encoding/scanning which are a basta
768.
▲
by
meaty
14y ago
Sounds like people shooting drones down with their shotguns. Some things were on the mark.
769.
▲
by
meaty
14y ago
Hmm quango resigns. And it's promoted as a bad thing?
770.
▲
by
meaty
14y ago
Yes I use keepassx as well now :)
771.
▲
by
meaty
14y ago
Either labour costs are cheaper (the Chinese are starting to want real money vs many unemployed people in the us) or they're getting a tax break somewhere or a subsidy.
772.
▲
by
meaty
14y ago
This. I'm a solution architect at a Microsoft gold partner (a loveless job). We are forever getting screwed. We've just managed to get through the windows workflow engine rewrite that appeared in WF4 and deprecated a shit load of stuff stil
773.
▲
by
meaty
14y ago
Yes 100% this.
774.
▲
by
meaty
14y ago
Annoyingly I did this once and then promptly forgot the port number, resulting in nmap time :(
775.
▲
by
meaty
14y ago
He is correct on a lot of things if you throw popular opinion out of the window and consider what is best for you* * human nature is inherently self-destructive. He seems to identify that it doesn't have to be.
776.
▲
by
meaty
14y ago
Good suggestion, but unfortunately I have banana fingers so can't use qwerty phones that well. T9 still works for me though :)
777.
▲
by
meaty
14y ago
Our biggest mailbox is 200mb. Pretty much everything gets deleted or moved out of the mail system. It's not a file system. If you think it is, you're doing it wrong. Backups: tar and gzip daily, then scp to a friend's server in another coun
778.
▲
by
meaty
14y ago
TBH I've only seen a couple of "rotten" machines in the last few years. It's not all that bad. Regarding WP, it's a fucking mess. I've had two handsets now and while it's a good product at the core, some of the decisions (such as isolated s
779.
▲
by
meaty
14y ago
here's the best one which respects the user, works well and is extremely stable: debian.
780.
▲
by
meaty
14y ago
I'm a little surprised at how accepting people are here relating to the Amazon feature in Ubuntu. I would assume that there are a lot of critical thinkers here as comp sci people tend to have that mentality. Possibly not! The usual excuses
More ›