Yeah, definitely don't take a pay cut for a lateral move - it's actually harder to get a raise shortly after your move than had you not moved and don't forget that it's generally costlier and riskier for them to hire someone from outside.

Does your current department know what's going on?

Yeah I will if it's low for sure. Trying to decide what to do if it's dead even or a little high.

My first time around with this was when I converted from contractor to permanent. I was so surprised at how high the offer was I didn't ask for any more. Finally I figured out they were basically giving me $5k on the table so I would feel like I was negotiating. I called the HR guy back and told him I wanted the $5k. He said my boss already gave it to me.

I suck at negotiating basically.

Just my bosses. They aren't nuts about it but support me. In the long run I think they realize it could be a big asset to have me in this role.

We're slow now anyway and all the heavy lifting on our node/angular stack is done. We're in the much-less-fun maintenance phase where our projects are annoying enhancements which are hard, and risky, but don't incorporate any new technology - and our 20 or so dev/test environments hosted in rackspace break down on a daily basis.

Being given the 5k as a nice gesture is def a classy move and unfortunately does confirm being bad at negotiating.

Mason should cut a deal with us users to data mine posts and sell it to advertisers and become evil geniuses.

obv they should fix it, but this actually doesn't make sense. if you owned a site bringing you however many thousands 2p2 brings them each month, would you hand it over to some dude on craigslist so he could install SSL?

it's a situation where you stand too much to lose. your only options are to hire someone you know and trust to do it, or to pay out the nose for a large, reputable firm to do the work. so depending on their setup, it's something that conceivably cost thousands of dollars to have done. that still seems well worth it, though.

How much does a damn SSL cert cost these days? Freaking post in this forum and ask one of us to do it.

Let's Encrypt is free and only takes about 5 minutes.

Giving a random off craigslist root access to your server sounds like a great idea, too.

Again, there are a dozen of us in this forum I'm assuming who could do it and they would trust. I have no idea how to do it, but I'm sure I could figure it out if you put a gun to my head.

i obv don't know for sure, but this is probably not the same thing as figuring out how to do it on a fresh DO box or something.

it's the integration with the existing system that could be a pain: ensuring redirects always work, making sure everything works with a possibly custom install of vbulletin, they may have multiple servers and load balancing, etc.

not disagreeing with your sentiment, just saying the situation is possibly more complex than you'd think.

Out of all the things I ever did on a server, putting on SSL was the easiest, but that's also a single server on DO using Nginx, hosting 3 sites on the same server.

You just install the cert, tell the Nginx where the cert is, redirect port 80 to port 443, and start a cron job to re-register the cert every 2 months.

I don't think it would matter if it is vbulletin or not. The redirect happens at the proxy server level, which has nothing much to do with PHP. I don't know for sure how much it matters that you have N servers serving everything since I don't know if there are URLS needed for each server. Even so, I'm sure that DO, HostGator, or whoever the host is can do this for you for a reasonable price.

I'd rather not out the poster who may or may not still be the server admin, but if anyone knows, maybe ask her?

unless you have tons of devops experience, you just can't confidently infer like that from your own experience.

i mean, how many have times have you burned half a day doing some silly **** that should have taken five minutes? i know you've used CSS... is it that hard to imagine that something like this could be way harder than you think in some context that you know nothing about?

it would also involve some code changes too right? between that, updating the web server configs, and getting the cert issued is probably a day at least

I only confidently conferred the following:

Installing an SSL doesn't take very long. This is not a controversial thought because it is very fast, plus you have to restart your proxy servers to get the redirects to work. If installing SSL took all day, the entire web would be down on some "install SSL day."

The redirects are from port 80 to port 443, which has absolutely nothing to do with any application code.

Using a rando off craigslist is a terrible idea because you need to do the above with root access.

There is someone who does the system stuff. I'd suggest sending said poster a PM about it if you know who it is.

I did allow that, if this site is loaded across many servers, then there may be other problems. Since I don't know how to do this at all, I do not wish to speculate or speak on this situation.

However, I am pointing out that there are some misconceptions about how SSL works: it has nothing to do with vBulletin, PHP, or any other programming language that is used.

For a decently quick overview:

https://www.digitalocean.com/communi...n-ubuntu-14-04

One of my blog posts is featured in PostgreSQL Weekly.

The key thing is digital ocean articles giving you step by step instructions for specific OS versions running on their own servers.

Yes, setting stuff up is easy when that is available.

you're wrong, it's possible application level code is coupled to SSL. i've had things break myself. maybe it *shouldn't* be that way, but it *can* be.

link to blog post please

Reading this makes me very sad. I'm sorry.

http://coderedux.com/on-postgresql/upsert-bad-apples

Yes, I see the irony of not being https at the moment. For whatever reason, I forgot to do it in this site and can't change it now due to the traffic spike.

Yeah, so, not every server is sitting behind a proxy that will handle SSL for you. Maybe it should be, I suppose it depends.

If it were me, and there was a proxy, I would set up a 2nd proxy, get https working on it, and when I was confident, swap the proxies. This way there would be no down time and I could work on it as I wanted to. I do this kind of thing all the time when you Really Must work on a production system with no down time.

If there were no proxy, I'd add one, get https working on it, and then point the twoplustwo domain at it. Similarly, no interruption. Piece of cake, imo.

But I know zilch about the IT infrastructure behind 2+2, so, who knows. I also have no real idea what the staff is like, if there is any.

nice article, dave

Thanks, glad you liked it. Hopefully you got something new out of it. It was an expected discovery for me, to say the least.

It was quite a surprise to see my article in the email. I have no idea how they found it.

One recommendation as someone who is usually a user for this kind of thing: if you are going to trade using it, you'll probably find the numbers clearer if you change the convention so that "no" is 100 - the price. So you would see, for example, Hillary is 86c / 85c instead of 86c / 15c. Buying a "yes" contract and selling a "no" contract are financially equivalent (I think? Haven't used PredictIt) so it makes sense to price them in the same terms.

There won't be trading, the PI API doesn't allow for it, but on PredictIt you can buy both "yes" or "no" on a contract, with prices that usually add up to near 100 (but with poor market efficiency, they take a fee out of profits so what would usually be arb opportunities become unprofitable). The numbers in the "no" column are in fact what you would pay to take that side of the contract.

Some people may or may not data mine POG WW games for their W/V/S classifier already :P
Which is fairly straightforward up until the algorithm tuning because you have a ground truth for the initial dataset due to the knowledge of roles post game.

While we're talking about that...some time back I posted along the lines "I must be missing an obvious resource for human classification to create a gold standard for machine learning" which was about the harder case where you don't have said ground truth and basically only a human expert (or crowd) evaluation that substitutes for truth.
Best resource I found: "Natural Language Annotation for Machine Learning" (O'Reily book). I also waded through a bunch of computer linguistics papers (corpus annotation). I'll iterate through a couple of annotation rounds and then hopefully the classifier will do what it's supposed to
Fairly interesting overall so far.

So? Do it. The cost of the inevitable breaches will be higher. Pretty amateur hour to say "oh noes we can't use SSL because it can't be done". It's just a matter of money. I guess the user's security just isn't worth it to the good folks at 2p2.

I want to start a personal project I can dive into this summer to keep my mind engaged, but I have no idea what. I'm not so good at finishing books/tutorials and I have like 0 ideas. Any ideas where I can get ideas? Sorry for being so vague. I just feel like all my classmates and peers know so much more than me and i'm just kinda drifting through my classes without ever applying anything anywhere else.