Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
alnayyir
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
28 ms
·
61.
▲
by
alnayyir
15y ago
I think ted was likely a great deal smarter than the people we're dealing with now.
62.
▲
by
alnayyir
15y ago
Particularly if they want to be bought into plus.google.com's team.
63.
▲
by
alnayyir
15y ago
Prolly for the engineers like everybody else lately.
64.
▲
by
alnayyir
15y ago
Present concurrency is 0, it's just slow. It's not SQLite unless there is pathological behavior because I've demoed ideas to HN traffic with SQLite before and it performed better than a vanilla MySQL deployment. (Not saying this is advisabl
65.
▲
by
alnayyir
15y ago
The funny part is trac isn't that fast either. It's overall a superior tool for the job, and loads simpler in every respect, but I get a healthy amount of delay on each and every action I perform in most trac deployments. I keep meaning to
66.
▲
by
alnayyir
15y ago
I'll borrow one from C-land: Undefined behavior!
67.
▲
by
alnayyir
15y ago
Optimizing for the wrong problem.
68.
▲
by
alnayyir
15y ago
Gen X'ers bitching about being employed. I couldn't write better comedy if I wanted to.
69.
▲
by
alnayyir
15y ago
It's not. At all.
70.
▲
by
alnayyir
15y ago
I just saw this exact conversation a month ago. Serious Deja Vu.
71.
▲
by
alnayyir
15y ago
Stop wasting our time with this. Post it when it's done, divided into parts or not. I will read part 1 and only part 1 if that's the only one available at the time it was posted. I will not return to your site to look for that content again
72.
▲
by
alnayyir
15y ago
I walk around plugged into my iPhone, portability isn't the issue, perturbing my physical state is. If I have to pause to think to move around or have an alarm that does this for me, I'm going to lose flow instantly . Edit: Point being, I'
73.
▲
by
alnayyir
15y ago
The code writing isn't the hard part per se, it's simply that I work for a startup and have a lot of work to do. My best design decisions and code and productivity come out of flow. I'm into functional languages like you are, but my bac
74.
▲
by
alnayyir
15y ago
>get up and walk around as I chew through a problem in my head. Surely this doesn't constitute a break in "flow". I'm afraid it most certainly does for myself and other programmers I know. My "walking around and contemplating" is a comp
75.
▲
by
alnayyir
15y ago
But not anymore? :(
76.
▲
by
alnayyir
15y ago
Is it going to have the same/similar multi-replication for enterprise only limitation that Riak has or is this going to be funded in some other way?
77.
▲
by
alnayyir
15y ago
Not that many. The Russians were providing virtually the only capital support the antagonist had at the time so harassing them and shutting down their supply chain and income sources made a lot of sense. Our small team ops were effectivel
78.
▲
by
alnayyir
15y ago
Yeah right, last thing they need is people inquiring/crowding into their livelihood. They didn't like me very much anyway. I was a merc paid to fuck up their operation by enemies of one of their allies.
79.
▲
by
alnayyir
15y ago
http://www.gnu.org/software/emacs/elisp/html_node/Scope.html http://www.gnu.org/software/emacs/emacs-paper.html#SEC17 The docs and the man himself say it best.
80.
▲
by
alnayyir
15y ago
Couch is definitely a lot more honest about their limitations than mongo or riak, but my experiences make me hesitant to recommend it to anyone not intimately familiar with those limitations.
81.
▲
by
alnayyir
15y ago
>This ISK would be hard to launder back into real money Not true, I know of several Russians that live off the game with far less ISK.
82.
▲
by
alnayyir
15y ago
CouchDB is better'ish for larger datasets, but not for arbitrary scaling. MapReduce in CouchDB requires dumb full-scans if you're not just refreshing an existing view. Arbitrarily large data is the exclusive domain of hadoop/hypertable/cass
83.
▲
by
alnayyir
15y ago
Do you have any idea how much stuff we've written in Elisp? It's. Not. Worth. It. You people don't even listen when we try to explain that lexical scope is unnecessary and potentially even a bad idea. Threads? Come on. :\
84.
▲
by
alnayyir
15y ago
As an avid elisper: No. Just...no. The lexical scope doesn't make sense for the kind of scripting done in Emacs and having a unified language (unlike vim with many people bolting their preferred language onto it) for the most part is a hug
85.
▲
by
alnayyir
15y ago
So when are you guys going to back up the openness propaganda and open source something of substance other than protobuf?
86.
▲
by
alnayyir
15y ago
I don't know, I don't even like how the startup industry currently works. It mostly looks like a wealth transfer from the wealthy to the children of yuppies to me. I don't see much business value actually being created.
87.
▲
by
alnayyir
15y ago
Acquisitions are about talent and not products more often than not with Google.
88.
▲
by
alnayyir
15y ago
You haven't qualified your statements yet, nor demonstrated you know much about either situation we're referencing.
89.
▲
by
alnayyir
15y ago
It hasn't occurred to how superficial that sounds, eh? You should read up on what Woz did to improve/perfect the design and efficiency of the Apple II.
90.
▲
by
alnayyir
15y ago
It's like a GROUP BY statement in a query. What are we compromising and not compromising on? Which values/properties/goals? What trade-offs are involved?
More ›