I bet if you take a look outside of social media, phone apps and web startups, you'll find the situation is a lot different. Granted, that excludes a lot of the hot companies that everyone hears about constantly...
My company is an early-stage high-tech startup in Austin and the only developer on my team under 40 is our front-end guy. It's not older because we are using ancient technology either... High-speed network processing in C, control plane and management code in Python, and a modern web-based management interface (HTML5, CSS, JavaScript, etc.)
Could just be a quirk, but it was similar in the last company I was in (network security product company). I suspect it is because it is embedded systems development, but maybe it is because the types of products we were/are building. Inline network appliances where you performance is critical and you can't bring the network down.
CS is overrated anyway. 10% should be CS, 30% should be Software Engineers, and 50% grunt work Programmers. All very different education.
I take it the other 10% should be in math?
Kind of hard to pause something the said they wanted to do. Which means they didn't even start it. Maybe notes on the back of a napkin. But that would be giving them to much credit.
Really? The 900 Mbps+ up and down I enjoy at my house from AT&T Gigapower is imaginary?
AT&T pausing their gigabit rollout when the President announces that he wants to make broadband a utility is completely reasonable. They have no idea what is going to happen, so it is hard to justify continuing to spend $$$ with the network upgrades.
Now, that's COMPLETELY different than not rate-shaping different types of traffic or trying to double-dip by charging both the sender and the receiver for traffic. Pretty much all of the ISPs are being butt-nuggets on that one.
Sounds like the real problem is that you are unwilling to relocate. Putting your company somewhere where the cost of living is high and there's a shortage of talent seems to be very popular, but difficult to understand.
One of the main things that drives this is how funding works. It's amazing how difficult it is to get funding if it requires the VCs to travel. Certainly a significant hurdle even for places like Austin where you have a decent-sized high-tech community.
Since there are already WAY more companies than they'll ever fund just down the street, it's hard to blame the VCs for not wanting to get on a plane constantly. Founders know this, so guess where they tend to start their companies?
Of course, while they are more likely to get funding, they also have major issues attracting and keeping talent. Catch-22
Just like we don't spank kids anymore because it's pointless and counterproductive, we should also stop "spanking" non-violent offenders but put them to good use instead.
Not sure which "we" you mean, but there are plenty of parents that spank kids. It isn't pointless or counterproductive.
For the relatively simple network settings (port config, user admin, etc.), CLI is fast and easy. Hard to imagine trying to deal with complex policy configuration from the CLI. Do you end up switching to the GUI for that?
Good UX design applies equally to CLI as well as GUI, which I'm guessing you actually do give a flying flip about.
Yes, which is why I asked about UX design instead of "best GUI"
For all the folks writing up the HTML code that goes into these things: use relative URLS!
Do not put the hostname (or IP address) of the device in any of the HTML. Us IT folks sometimes need to go through proxies (and SSH tunnels) to get to these devices (which are often on isolated "management" VLANs/networks). Simply put "/network/settings" instead of "http://mydevice/network/settings" in any [a href] links (or [img] or CSS references).
If the link in the HTML has "10.10.20.45" or "netdev01.mgt.example.com" in it, but my browser is actually connecting to "localhost" (because I have to do a SSH double-hop with forwarding), I'm going to think really evil thoughts about whomever wrote the HTML generator. I do not have to want to start editing my
Great point and something easy to miss during the mayhem of implementation of a new product.
Not at every entry point, security should be a serious consideration on every device. Work on the assumption that everything is directly exposed to the internet and start from there. Trying to only monitor the entry points is the problem, if anything makes it past your entry points then it could have free reign over everything inside.
When I said "entry point", I didn't mean the perimeter. I meant at every single connection to the network... the RJ45 you plug into in your cube, the wireless AP you connect to with your laptop or smartphone, the vNIC in your virtual server, etc.
Personally I think BYOD is a disaster waiting to happen, but whatever.
If you don't trust any device connecting to the network and IF you are able to apply appropriate security inspection to all of the traffic, does BYOD actually matter? From the security perspective, I'm not sure it does. That said, I tend to agree with you from multiple other concerns: IP protection, compliance, backups, support, etc.
My thinking on this is a bit different, and boils down to this principle: There's still a perimeter, but most of the office is outside of the perimeter.
So what do you include in your version of a perimeter?
All seems condemned in the long run to approximate a state akin to Gaussian noise. -- James Martin
