Forgot your password?

Comment: Re: Unified Experience Across Devices (Score 1) 394

by guruevi (#48030731) Attached to: Microsoft Announces Windows 10

Windows 9x-ME was really Windows 4 all along. 2000 was version 5, XP-10 is version 6.

Most windows versions suck, regardless of version numbering. They suck less when it comes around to having a third service pack but they're still miles behind a real OS. (I haven't used windows computers at home since windows 3.11)

Comment: Re:In US, restrictions based on finite RF frequenc (Score 1) 108

No, it means investing in better antenna equipment, I can get gigabit speeds on an unregulated frequency, a regulated frequency should be much easier. Japan has 100Mbps to individual mobile devices, setting up P2P wireless links is even easier. Even so, the country has paid said regulatory fees to ensure wired access to everyone.

Comment: Re: Huh? (Score 1) 125

Other than the contract wasn't for building a launcher in the first place, hence why the remarks about experience in building a launcher is irrelevant. The contract was for building a spacecraft that would sit on top of a launch vehicle. In the case of SNC, they were using the services of United Launch Alliance, a company who has experience in launching stuff into orbit. ULA has been putting stuff into orbit (at least their parent companies) since the 1950's. Is that enough experience?

It helps to read the fine print.

Comment: Re:Building a satellite is really (Score 1) 125

The contract is not about launching the spacecraft, it is about building them and having the work in space.

Besides, SNC is going to be launched on the same vehicle that Boeing is using. The Atlas V. The only difference is that the Dream Chaser could also be launched on an Orbital Antares rocket or the Falcon 9 as well (at least it is being designed to fly on multiple launchers).

That isn't even a consideration for why SNC lost the bid.

Comment: Re:So offer a cost effective replacement (Score 1) 185

by JWSmythe (#48007299) Attached to: Security Collapse In the HTTPS Market

I really liked PayPal's solution for limiting risk when paying sites that didn't support PayPal. Their Virtual Debit Card product was great. I could provide whatever information I wanted, restrict the virtual card to exactly the amount of the transaction, and optionally allow it for recurring transactions. They were awesome, especially when purchasing from small companies with very little information about if they were legitimate or not.

PayPal if nice and all, but plenty of people fall for the common traps, like variations on the domain name which are phisher traps.

People here were generally better at avoiding scams, but that doesn't help the > 90% of the population who never check.

Comment: Re:min install (Score 1) 221

by MikeBabcock (#48007087) Attached to: Outlining Thin Linux

Aside from someone else already pointing out that you want to use different tools, that's exactly my point -- their minimal install is truly minimal -- there's no need to roll your own at all.

My basic install procedure is a CentOS minimal with a quick shell script that installs the packages and configs I need on top of that on a per-client basis.

Comment: Re:min install (Score 1) 221

by MikeBabcock (#48007079) Attached to: Outlining Thin Linux

If you're putting together your own optimized small footprint installs, you're not a sysadmin anymore, you're a distro maker. I spend way too much time actually administering working machines to be bothered to do someone else's job as well.

CentOS does a fantastic job of maintaining their minimal install for me (and anyone else who wishes to use it), what possible advantage is there to me putting together something else (not to mention learning a new filesystem and config layout for no reason).

Comment: Re:Samsung stockholder applause? (Score 3, Insightful) 88

by Teancum (#47998763) Attached to: John Carmack's Oculus Connect Keynote Probably Had Samsung Cringing

With the money Mr. Carmack earned from the sale of Oculus, do you think he cares? He has suggested that he wouldn't mind simply going back to running his spaceship company, so can Samsung give him a good reason to do that?

This definitely sounds like somebody who doesn't give a damn.

Comment: Re:In US, restrictions based on finite RF frequenc (Score 1) 108

by guruevi (#47990143) Attached to: Not Just Netflix: Google Challenges Canada's Power To Regulate Online Video

The 'spectrum' or bandwidth of the Internet is virtually unlimited though. You just need to put in bigger pipes and even the smallest of the pipes you can currently get at an IX (1Gbps) can easily carry 1000 simultaneous viewers.

The ISP's only have exclusive rights to the last mile because we (the people) let them. For the most part, "the people" paid over and over again for this last mile as well as all the other miles (both phone and cable) through regulatory fees but either is being monopolized by a single provider. There is no technical reason that several providers couldn't offer you the 'last mile' connection. It's being done in several European countries where you have a pick of providers to offer you the last mile.

Comment: Re:Read Slashdot (Score 2) 471

by JWSmythe (#47981319) Attached to: Ask Slashdot: Finding a Job After Completing Computer Science Ph.D?

The rejections you got may not have been because you didn't know a specific answer to a very technical question.

Something I've come across in the past is something similar. It's not knowing the specific answer. Sometimes it's knowing what specific answer *they* want.

For example, "How can you change the IP on a current RHEL or CentOS box".

There are a bunch of right answers.

  • edit the appropriate /etc/sysconfig/network-scripts/ifcfg-eth* file.
  • system-config-network
  • /usr/sbin/system-config-network
  • use ifconfig directly (not durable through a reboot, but ...)
  • change the static entry on the dhcp server for that network interface
  • modify it in cfengine, and wait for it to update.

... and those could all be wrong. That particular shop may say "We don't trust ifcfg-eth*", "system-config-network mangles the file format", or even "we don't use those files, we use /etc/rc.d/rc.inet1, that the old admin 10 years ago wrote". It could even start with "fill out the production change review forms, and submit them to the change review committee".

Some places insist that you use the full path to scripts, in case someone else put one farther up in your path (like /bin/). Some don't allow sbin to be in the user's path at all. And of course, if you failed to say "use sudo", you're one of those renegade admins who thinks they can run commands as root. Not knowing *their* method, even though you've never worked for them, is enough to fail an interview.

When I've been interviewing people, I don't work from a hard set of answers. If the interviewee comes close enough, they got it right. If they gave the "system-config-network" answer, I'd just ask "Do you know what files that modifies related to IPs?"

I've interviewed with Google a few times. One of the questions they asked was "How does telnet work?" I answered, and the interviewer asked me the question again. I gave the brief description, the detailed description, all the way down to the opening of sockets and how TCP works. Finally I just had to tell him, "I'm not sure what you're looking for in the answer. Can you please clarify the question?" He didn't. I don't know if that was a pass, fail, or just a stress question.

Never test for an error condition you don't know how to handle. -- Steinbach