THIS!

Also I use spotlight to open everything. Hit command + space bar then type in the program you want, boooom

please disregard the massive brainfart of my last post

ready for a blown mind? Alfred2.

I asked my buddy who works at the new company. He said no framework. He also said the PHP code is a bit chaotic. Should be interesting...

Well, good luck. I'm sure you'll learn a lot of cool stuff either way.

I just started looking for another job after moving. I am tempted to just start my own company.

I have 0 clue about OSX (whatever came with the G4 Powerbooks was the last I tried and it felt really odd but that was when you had to install a seperate X to run X stuff). I'm assuming "any program you want" means "any program that is on the path"?

Why not just have a terminal open at all times and type it in there (or I guess alias the terminal to command+space)

Guess the advanced stuff also uses indexing+some sort of scripting to let you do stuff like "play songname" and the like?

[I'm generally a very non-power user when it comes to the GUI and I feel like I'm missing out on some cool stuff]

That's not the only thing Alfred does, and Spotlight remembers the names of app better than you can I dare say.

I actually think the case of "forgetting stuff" that comes up most for me is forgetting a "sudo" and thus for me "sudo !!" = greatest thing ever :P

Sounds like someone got a job at Facebook!

It's his first job and he's a self learner under 2 years. What kind of job could he expect to get that doesn't require running. I'd program PHP sans framework any day before taking a Drupal or Magento job.

God I'm so ****ing sick of recruiters. I'm still trying to wrap my head around the best way to deal with them on the $ side. I'm thinking about responding to the inevitable "What is your salary range?" question with "Well I'd like to make 80% of what you bill the client. What is that, exactly?". I wonder how that'd go.

You hiring? If it's for a small business I'd just avoid them entirely.

Wondering if anyone thinks this is the best solution:

I have a db that stores transactions. It's highly important I don't lose ANY records under any circumstances, so daily/hourly backups aren't a good enough solution.

The important data is annoynmous and non useful to any attacker (it's a reference ID, and a Bitcoin address to deliver Bitcoins to when the order is fullfiled)

Is the best way to ensure I don't lose records at any stage simply to send an email to a dormant email account containing the important data each time a transaction occurs?

Bit of a noob question, I use SQL server and am not sure if there's any features to address this problem specifically. Alternatively I suppose I could have another DB on a different network that only has INSERT permissions or something like that.

the above solution assumes that sending an email and storing it in an email account has a lower failure probability than storing a record in a database. that seems unlikely to me.

why not just have a redundant database on another server that gets written to any time the main one does? i'm sure there's a better/more standard solution even than that, but emailing seems like a poor one.

The redundant database on another server requires another server, and also in the event the primary server is attacked, the other database is now at risk of being manipulated/destroyed.

so does the email solution. it's just the email server is hosted for you. are you worried about cost? servers are so cheap these days i don't see how that's an issue

how is this different from the email solution? if you give insert only privileges to the main server, it's exactly the same situation. if main server is attacked, attacker can send you fake emails or fake inserts.

i suspect both of these are the wrong way to think about it. i just think email is more wrong

What are you trying to protect the data from? Accidental loss is pretty easy, I'm almost certain there is a level of logging that would allow you to rebuild the db from the last snapshot in case of hardware failure. Phantom hackers, however, always have more resources than you do...

I'm building a new startup which I'll hopefully be launching the MVP for in a week or so. I'm under the assumption that Bitcoin related transfer sites attract a lot of hackers.

I'm trying to protect myself from manipulation or destruction of data by having a read only ledger than I can refer to. For example if a smart hacker somehow managed to get access to my database they might change the addresses to send Bitcoins for each order to addresses in their control.

If I have a separate server with a database storing these records, even if it's only got write-to permissions, I think it's plausible the attacker may be able to compromise that system in a similar fashion that they compromised the primary server.

Email records seems a bit hacky, but it's the cheapest, easiest and possibly safer as exploits used to compromise the live server probably can't be used to compromise the email records. Even if email records were compromised, manipulation of emails probably isn't possible, the attacker could probably only delete them.

Instead of sending an email, can't you just write to a local file and then regularly rotate those files to something like s3?

Edit: Ah, I see that you're probably worried about someone compromising the contents of your local file. There are probably easy ways to secure that though and it would mean that a hacker would have to do two totally separate attacks (compromise DB and local file) that worked in sync to actually modify records.

Storing files on S3 looks like a good solution actually.

I'm writing a Java library that has to work with multiple versions of an open source Java project. Unfortunately the project has an Interface I have to implement in my code that is different in the different versions.

Anyone know of a good way to handle this? Java doesn't like having unimplemented methods or having methods declared as @Override when they aren't actually overriding a method.

My best idea so far is to have separate files for each version of the interface and specify at compile time what version to use. Seems crappy though.

Edit: New best idea - create an interface in my library that extends the interface from the open source project and declares all methods that aren't consistent across versions then I implement that interface. It makes writing my code a bit harder since I have to know which methods in that interface I can't depend on - but I basically need to do that anyway.

you are essentially just describing the facade pattern, no? that seems like the right fit here, if i understand you correctly.

Sort of. But the issue isn't so much that I want to hide methods/complexity from the parent project. It's that I want to implement an interface from it without requiring a version of my library for each and every version of the parent project.

Yea, lol. I still think it's a good opportunity for me. They're basically willing to train me for 6 months on the job which I think could be incredibly valuable.