He was asking about languages and frameworks and developing their web application for the future.

If they're redeveloping from scratch in order to future proof things, it's unlikely that they're going to be wanting to produce the exact same HTML that they always have. A framework to help you develop better HTML easier is something they should be considering at the same time.

It would really suck to rebuild your application, get to the end, and then decide it's time to make it more mobile friendly, at which point they realize they have to do another redesign because they didn't think about how their user interfaces were going to break down.

If they need to ask which server side language they should use, then they almost certainly need to have these things pointed out to them as well. If it turns out that they didn't need any of that pointed out to them, then they can easily disregard the extra information.

Globals have been disabled by default in PHP for a very long time.

"register globals" which allowed post and get parameters to be automatically registered in the global scope was defaulted off a long time ago, and in newer versions of php (5.4+) is not even an option any more. This is historically the feature that got a lot of bad programmers in trouble.

Almost every language has a way for functions to access global scope variables, and PHP is not exception, but to do so now, you have to specify exactly what you're going to access by doing it through _GLOBALS or by calling "global " inside your function.

The default scope for variables is to have no globals, and to direct you toward a more OO programming style. You can still shoot yourself in the foot, just like every other programming language, but you have to at least try a little to do it.

PHP of old used to make it very easy to write applications with large security holes, but newer versions do a much better job of preventing developer's tendancies to shoot themselves in the foot.

I think it will be a very viable choice for web applications for the next 10 years or more.

There are a number of frameworks written in PHP that are pretty good as well. For my current project though, I've chosen to write a framework that is geared toward exactly what that project needed. I did choose to use an HTML framework to aid in the UI creation and standardization. For my project I chose "Foundation", but there are a lot of other good ones as well.. If your application has a requirement of being mobile device friendly (is there anything that doesn't?) then I would highly recommend a 12 column HTML framework.. If you don't know why a 12 column framework is the way to go, Google it, there are plenty of write ups.

About two years ago, a community college near me had the exact same thing happen. I don't know the excruciating details, but the basics were the same SCCM wiped out all of the servers that it was used to manage..

I didn't work for the college then (I do now), but I did know a few people that did at the time. The person that triggered it is still there. From what I understand what he was doing and the way he went about it, although in hindsight was dangerous, wasn't a really reckless thing.

Our campus is less than 30 minutes drive to Microsoft's main campus, and there was a lot of pressure for us to use their systems. I think the college paid the price for caving to that pressure. Sure, there are other factors involved here as well... A careless employee, an unintuitive result from an interface/script, poor safety mechanisms in both policy and the product, poor design by both the vendor and the college..

From what I understand, one of the most devastating aspects when it came to recovery, was that the server that held backups (Microsoft's data protection manager of course) was wiped out as well.. I think in this particular incident, only system drives were annihilated, so if a server had a "D" drive or other volumes, it was still there, it was just a useless lump sitting on a server with no OS for a while at first.

Having never heard a similar story with any other software product, I'm left believing that SCCM and it's deficiencies are at least partially to blame, and given what I know about the person that caused it here, I'd say that it's a pretty respectable bit of the blame that should be left on SCCM.

Someone realized pretty quickly what was going on (not the person that caused it), and pulled the plug on the process somehow or our college would have been even more devastated. As it stood, it was still pretty bad.. Probably only about 25% of the full destructive power of the mess as averted.

Google docs will let you connect multiple people to the same spreadsheet at the same time..

It works pretty well too... as long as the slightly chaotic editing that this creates is OK (like you don't need multi-cell/multi-sheet locking to keep people out of each other's business)..

Do you read XKCD, and if so, what do you think about the accuracy Randall Munroe's typical analysis?

It's more like getting a thimble full of water from the ocean, seeing a bunch of fish in it, and then going where are all the fish in the ocean.

I used to work there (on that line for a while), and one of the jobs was to beat them up a bit before they went out the door, just to make sure they could take it.. (We were careful not to scuff them up, but did need to subject them to a couple of impacts in each direction as part of the final testing).

Note, when he took it apart in the video, he very likely *did* make it go out of spec at that point.. It's normally just the high voltage that goes out of spec, but would normally only mean that you got a reading of 1007 VAC instead of 1000 VAC.. Still somewhat close..

He should send it back for recalibration after his adventure..

OK, first, you *can not* mine 10 or so bitcoins. There were no mining pools at first, and that is the only way people mine fewer coins.. And that's not really even correct.. Mining pools mine 25 bitcoins these days, and then share them with their members.. What we're talking about is directly mining coins here, which got mine 50 coins at a time for the first four years or so.

Also, IIRC, most of these coins are held by just a few addresses, not spread among 100,000. The entire population of the bitcoin community was probably less than a few thousand people during the first year.. During the first months it was more like 20 or 30.... maybe less..

I think you mean until he caches them into United States fun bucks..

...you wrote the IDE yourself :)

I believe that Xen is probably the most used VM/hypervisor on the planet..

Amazon and EC2 have deployed a pretty fair number of them with a lot of success (like all of their VM's).

I was at a Chuckecheese with the kids for one of their friends birthday parties when one of the machines freaked out...

It was a photobooth that took your picture, and then made a sketch like version of your picture and printed it out for you..

When the employee came to reset it, I got to see either Redhat or Cent boot up.. Somewhere I've got a picture..

Even it's ability to chirp loudly when a drive fails?

I think you've used some pretty lame raid controllers.

How about it's ability to not waste CPU?

ZFS is good... great even, but (irony intentional) absolute statements are always wrong!

I've used them together. Seems to work just fine.. Just don't let ZFS know that there's more than 1 drive. You can't have them both trying to manage the redundant storage.

ZFS has some great features besides it's redundant storage. You can get them from other filesystems too though I suppose, but I like snapshots built into the filesystem. It *is* overkill to have the filesystem doing checksums and the raid card detecting errors as well, but that's why this is the paranoia setup... Not really looking for the performance king..

ZFS certainly isn't necessary though, if you've got hardware raid.