I struggle to compose anything meaningful on twitter within the character limit, so I applaud this move. Not that anyone reads me on twitter anyway. But my replies sometimes get replies.

https://developer.mozilla.org/en-US/...MLInputElement

ctr+f "defaultChecked"

It's another dom element property. React just maps props to dom nodes unless it is marked as having special functionality/meaning,

https://github.com/facebook/react/bl...ations.js#L192

^^ thanks interesting.

Didn't Twitter completely screw their developer ecosystem over at some point? I don't really follow this stuff but I remember hearing a presentation from someone that had a good startup going that got really screwed by Twitter closing down/changing a bunch of stuff.

I like the change, but I don't tweet much. I just use to hate when someone would ask me a technical question and I had to condense a complicated reply into multiple small tweets.

Isn't that nearly universal?

Twitter: loads of people talking in one-line text speak to a bunch of people who aren't listening, responding with their own one-line mumbling no one else is bothering to read.

I suspect in the near future, "tweeting" will be used to describe two old deaf guys yelling at each other at the bar, talking about two entirely different topics but still enjoying the conversation. For now, I think of Twitter as the Zombie Apocalypse, both real and virtual.

If people are following their spouses and best friends and aren't even reading that stuff, why are they going to read and share tweets from McDonald's?

Hi guys, scaling question:

Today I added a line of code that basically updates the modified time for a large join table (100 mil rows) at an API endpoint. The table has only 4-5 fields, and I make the SELECT on primary key. This endpoint gets called every 15 seconds. After a few hours or so, our entire site was at a crawl due to these updates.

Since our company's release process is quite open, I'm trying to avoid pushing these types of changes in the future. The thing is, I honestly thought if the servers could handle the amount of requests, then 1 UPDATE per request wouldn't be that expensive (since it just locks row). Am I just thinking about this the wrong way? How should I approach these types of problems?

Probably better for dba.stackexchange, but at least say what database system you are using.

Not sure what a join table is. Do you mean intersection table?

You can't really chose if a statement is row-locking or table-locking. That is up the the database engine.

Writes can be very slow, but I suspect you have something else going on, like cascading foreign keys or something else that is causing a slow down. Are the timestamps cascading to other tables? It really doesn't make sense that this would take 15 seconds.

Why wouldn't you use a trigger to run this operation? There is no good reason, speed-wise, to make multiple calls between the code and the database to update, yeah?

Something is confusing about the description. How many rows are you updating each time?

Yea, I'm pretty sure there were numerous people who went into work and their companies were over.

First they introduced changes to their API where guidelines became requirements, and throttled the amount of users you could have before Twitter could choose to support you or not.

Then they more recently took away firehose partners, though I have a suspicion some still exist.

http://mashable.com/2012/08/16/twitt.../#KJcEpPR_qEqF

Roger, maybe post the actual code, appropriately anonymized.

2nd POC-type project in a row at work has come to a grinding halt due to lack of network connectivity between disparate groups (within the same giant corporation). I feel like we need a manager of network connectivity, where literally all that person ever does is try to push through new network connections.

My work is planning on making us do all of our work using VDI over a network connection to secure data center in Las Vegas. We are taking bets on how long it will last before they figure out it is a terrible idea.

Our offshore people work over a VDI. It takes like 30 seconds after they click something for the page to react. Seems to be a feature so they can bill more hours.

I still don't get how incredibly awful and slow lync is vs webex re: screen sharing.

Meerkat?

http://www.theverge.com/2015/3/13/82...t-as-sxsw-gets

I got SICP for Christmas and have been working through it. Downloaded the MIT/GNU scheme package which comes with a program called Edwin that is an editor/interpreter Emacs variant. This is a bit of a pain to me since I've never used Vim/Emacs etc.

DaveT, is this what you used when you studied the book? Any other scheme interpreters you can recommend?

https://racket-lang.org/

May also want to look into: http://stackoverflow.com/questions/1...p-in-dr-racket

Thanks!

Re: Database updates

Sorry it was as Dave expected, it wasn't just the update, there was some cascading. We're using Django ORM, so I think I might have overlooked a signal somewhere.

Yeah, I used Racket + Neil's SICP plugins as well. Seems like the plugin is no longer maintained, but there are some items on github available as well.

http://www.neilvandyke.org/racket-sicp/

I would say that, long-term, it is worth putting forth the effort to learn Emacs or Edwin. I know it is scary and a PITA right now, but I still have memories of using Notepad++ in my early days of Clojure. At some point, you'll have to bite the bullet and learn emacs along with some language in the future. I don't have an argument either way, but food for thought. I will say that learning Emacs is among the single most important thing I learned.

I also recall the Racket IDE had a ton of issues as well. When it worked, it was pretty good, but when it failed, it was a total catastrophe. I'm guessing it improved quite a bit by now. My main complaints are that it was very slow, bound the crashing, and just generally a displeasure to use.

If you are ready to take a shallow dive, Simple crash-course:

Emacs is mnemonic. Basic movement commands:
F = forward
B = Backward
U = Up
N = Next

Emacs also supplies what I call "minor chords" with CTL and "power chords" with ALT:
To move forward one letter: CTL + F
To move forward on Word: ALT + F

By convention, we write "C" for "CTL" and "M" for "ALT", so:

Other examples:
K = kill line
C + K: kill one line
M + K: Kill paragraph

D = delete
C + D: delete letter
M + D: delete word

With all that said!

For right now, you can use the arrow keys, and then execute each statement with C-x C-e (you can hold CTL down for this) and M-o for evaluating the whole buffer.

While "switch buffer" is probably C-x o, you can also use the mouse for that as well.

So, really, just need to know how to find and create files:
Open edwin:
C-x C-f (you can hold down CTL here)
look at the mini-buffer and find the file you are looking for. You can use tab-completion and it works just like the command line.

press enter on the file.

If you are creating a new file, just type a name like "problem1.scm", then press enter. Probably have to press enter again for "ok".

Then you need to know how to evaluate:
You can then evaluate just one expression or the whole buffer. I'd suggest KISS here and creating one file for each assignment, then evaluating the buffer or the code part you want. C-x C-e will evaluate whatever is to the left of the cursor. M-o will evaluate the entire buffer.

All you really need to know is how to open and create a file, execute a statement, and stop an execution.

As you get further, you'll end up using "M-x" quite often. Using "M-x" is, IMO, the hardest thing to really learn how to use, and not strictly mandatory yet. The rest is just a matter of learning some basic things as you need them.

You probably want to know copy-paste:
C-space (start mark at cursor)
move forward or back to highlight what you want to highlight (use arrow keys or movement combos here)
M-w (copy) or C-w (cut)
move to where you want to paste
C-y (paste, but Emacs users say "yank")

And if you end up with an endless loop and need to stop evaluation stat:
C-g
(or just click the "x" in the top left corner and kill Edwin entirely)

Thanks for the tips. I'll give it another go. I was travelling over the holidays, so it should be easier now that I am back home and can use two screens again.

lol technology

https://twitter.com/williamnyy23/sta...32975725395970

Emacs is useful but I wouldn't say the same for Edwin and MIT Scheme, which aren't really maintained and almost obsolete. Racket is a far more complete development environment that has kept up with Scheme standards, not to mention bindings/libraries useful for modern development and you don't have to use its IDE - you can always run your programs from Emacs.

Edit: I guess if you're just trying to get through SICP, it doesn't really matter, but in that case Racket IDE > Edwin + MIT Scheme > Emacs + Racket

FWIW, I took SICP as a course in school with content translated to another more obscure language (with an object system built-in), using an interpreter/REPL implemented as a Java applet. I maintained the runtime (just fixing minor bugs) the following term and it was indeed terribly implemented but it didn't really matter for the course.

If he attempts to use vanilla Racket Scheme, he will run into quite a few issues as many things are not compatible, and I think it makes as much sense to use Racket for SICP as it makes sense to use Clojure in SICP. Some things don't really matter, but Racket doesn't allow you to do many things MIT Scheme allows you to do. Trade offs here and there, but they felt super important to me when I was going through that book, plus it is important, IMO, to not have the fall-back to built-in items. Not sure how I'd feel about those differences today.

Edwin is basically Emacs with the same key-bindings without much of power of Emacs. I think Racket IDE has quite a few issues and it isn't a very good IDE outside of paren-matching and a run button (and not very good at either, to be honest). Racket is slow, bound to crash, and really is a displeasure to use. I don't think much of it is transferable to a more common IDE. Edwin at least is transferable and that knowledge can be built on. I think the long-term gains will way outshine the short-term irritation with Edwin. I can't make that argument for Racket. I'm just giving what I wish someone told me way be when.

It is too bad the Van Dyke plugin is no longer maintained, which did a very good job of allowing you to break the rules of Racket.

In any case, just be sure to watch the 1985 videos. There is no replacing that.

With the caveat: If I had a real-life teacher, I'd probably feel a bit different.

So I'm coming up on 1 full year of employment as a software engineer at this company. There has been no mention of any salary increase and the holidays has passed without any end of year bonus even though business has been good this year. Should i be concerned?

In the interview last year, my boss has said he has never not gotten a bonus and there's a 3-5% raise each year. Of course for him, it's probably true as the owner seems to love him.

Fyi, they see me as a recent college graduate who just started working. But actually, i'm a former poker pro who had to reinvent myself to be able to get hired. Since this is my first "real" job, I'm just curious of typical salary increases going forward. Maybe my expectations are completely off.