Re: high school
You: “30 years ago, not relevant”
Recruiter: “we weren’t going to hire you because of your age, but now we’re going to not hire you because you didn’t humor us to answer the question.”
79 publicly visible posts • joined 15 Jun 2009
> This is correct. Modern DCs for AWS / MS etc are "lights out" facilities that contain very very few humans.
My guess is there are probably 20-30 humans needed at a data center once its up and running. You need people for:
* A couple of people receiving shipments
* A set of electricians on site most of the time just to keep the lights on, service generators, etc
* HVAC engineers
* A couple of network operators on-site
* A couple of people to rack servers, replace failed hardware, plug stuff in
* A couple of security guards
* Probably a manager or two
None of these need any specific high-tech skills, but there's a non-trivial amount of labor that has to happen.
> Narayanan said having 13 machines across eight libraries saves the NLB approximately 3,500 hours of labour a year, covering about 50,000 books overnight.
Assuming these machines only work at night and a night goes from 12am to 8am, this is 104 hours / night, or 37,960 hours per year. And if that's equal to 3,500 human hours, then one hour of robot time is equal to 0.09 hours of human time.
It'd be interesting to know what goes into computing that (are they subtracting maintenance hours from the hours saved?).
I think you're right - PostgreSQL is a much easier sell for new stuff than Db2. I could totally see IBM buying EDB - Keep existing Db2 customers happy, and make sure that if those existing customers are tempted to stray from Db2 to open platforms, you can upgrade their support level to include PostgreSQL support and retain them.
If you choose PostgreSQL and you want support, you have a couple of options - Percona, EnterpriseDB to name two of the bigger ones. You can have them give compete with each other for your business. If you don't like the support you're getting, you can switch.
If you use DB2 and you don't like the support or terms IBM is giving you, you can think about the life choices that got you to that place, but not much else.
If you're starting a new project, I can see no world in which I would choose DB2, Oracle, MSSQL over PostgreSQL. If it wasn't for PostgreSQL, then there's no way you'd choose DB2, Oracle or MSSQL over MySQL.
PostgreSQL and MySQL are open source and have enterprise grade 24/7 support available from multiple vendors. Who would voluntarily shackle themselves to a single, closed source platform for the foundational component of the stack which may be the most difficult to migrate to an alternative?
Moving Db2 to the cloud will protect it and give customers one less reason to migrate off of it, so that's a smart move. If IBM wanted Db2 to compete for new business, they'd have to open source it. Otherwise no one in their right mind would consider it for new development.
Relational data stores, compute infrastructure, servers - these are things that are provisioned and priced like utilities now.
Whereas before you needed a full-time DBA just to set things up and keep the lights on, now you can pay a monthly fee to have that provided with relatively well understood SLA's.
There's still a need for skilled workers - but they need to be able to join three or four utility systems together into one product. That's not a jack of all trades, its just a different specialization.
It's classic penny-wise, pound-foolish behavior.
* A $20 debt from an inactive account certainly costs more for them to collect (in AR staff time) than the amount of the debt.
* It generates ill will within a class (technology professionals) it is actively hiring from.
* It looks bad for a technology firm to say "gee whiz, we don't know how we overpaid you. Our record keeping and accounts payable need some better integration.
But aside from all of that... Aside from matters of perception and economic benefit, common sense and human decency dictate that you do not do this.
I bought a few overland products a few years back. They were cheap, had a ton of capacity for the price, worked 95% of the time (which really isn't great) and were cheap. Also they were cheap.
Kind of warms my heart to see that they're out there still making cheap stuff with tons of capacity. Not thrilled to see them move to China but quality wasn't why you bought them anyway.
I worked in a lab that did hard real time experiments and used DOS - running on modern hardware with e1000 adapters and some highly specialized software for recording inputs. No graphics, nothing too fancy but deterministic low latency requirements.
I was always impressed that Intel making and updating the e1000 drivers for dos. Shows that they had their software engineering under control that they could keep that commitment.
Why would I need a massively distributed ERP system, and why should SAP fear one?
Sure, cloud is the marketing term du jour, but "real cloud" isn't a panacea.
If you have a big enterprise, you have lots of uses for a system of record, and it won't need to scale out over a great huge cloud. You'll have plenty of uses for SQL databases too. SAP and Oracle are trembling all the way to the bank.
Many of the apps show use the built in maps.
Heres an example: use Zillow or Trulia apps to find a house for sale... iOS5, you can see train stations on the map, which is useful when assessing the location of a house to purchase. But as soon as you upgrade to iOS6, that train statio. No longer shows up. Doesnt sound like much, but compound it with everything else that uses the built in maps... non-trivial downgrade.
For the sake of perspective, I work at an American tech company, we have more H1-B workers than natives in my department.
I think H1-B, and more immigration is a good thing. I like that my company has a bunch of Asians (and me) working here rather than moving our department to Asia.
There may be an Apple TV, but it won't be for just watching TV. The TV watching part will be like an iPhone making calls. It will do so, acceptably, but its not anything to write home about.
If Apple does a TV, it will have Siri, recognize voices from different members of the family, and be more iPad than TV. Interface will primarily be voice, but you can use your iOS device too.
For most people most of the time, it will be the most expensive picture frame they ever had. But there will be Netflix, iTunes, Facetime, and somehow Angry Birds.
And there's one more thing... You can watch TV on it! (just plug it in to your set top box and use that box's remote.)
The iPad is the standard. They've been out for a while, I know people who have them. I have used it extensively but I have not purchased one. Its like the iPhone, which I think is pretty useful.
I may have seen another tablet on the train. Tried a [xoom|galaxy|playbook|webos tablet] in BestBuy, but it was a pain because the security cord was too short and the wireless network was slow, and the interface not intuitive enough to pick it up well enough in the store.
They all cost about as much or more than an iPad. I might like Android better than iOS if I gave it enough time. But I'm busy.
What it comes down to is this:
I know I like the iPad. Don't love it, but its pretty nice. i know I might like a [xoom|galaxy|etc]. But I don't know for sure. Might love it, might regret not getting an iPad.
If I wanted to do more couch surfing, why would I buy anything other than an iPad? Its a know quantity and the decision is easy.
If there was some killer app for another one then I'd consider it. But they're are all focused on copying Apple, and they haven't done a great job of it yet.
But its also pretty clear that journalists are a bunch of hacks that don't bother to fact check the material they report!
(Wait, did I just comment on an article without following the links it contains? Is it possible el Reg is pulling one over on me to demonstrate something about its readers?)
"I do not know who has given these documents to the media, but the purpose can only be one thing - trying to make Julian look bad."
I mean, if you replace Julian with United States, I'd imagine you could find a US government official who said the same thing.
The one thing that bugs me about this - why leak it to the Guardian? What if there were some sort of website out there where any could anonymously post stuff? Sort of like Wikipedia, but with information that had somehow "leaked"? I'd call it LeaksWiki.
As I understand it, the only way Sun (and now Oracle) makes any money off of Java is buy licensing it.
Everything is open, except the test kits (TCK's) that you need to get your version of Java certified. Those you have to pay Oracle for. That means you can create your own java implementation without paying Oracle. And Oracle is suing to say that you can't create your own java implementation and call it something else.
If the TCK's were open source, and Harmony went forward, how would Oracle be able to make money off of Java? Harmony would be a fully open end run. Everybody that is paying wants this. Everybody that is getting paid (Sun, then Oracle) don't want this.
When Oracle didn't own Java, they felt differently. After they spent a ton of dough to buy Java, they are trying to continue making money from Java. And why wouldn't they?
A lot has been made of the sample size in the comments. I wonder if it was really a random sample or not.
Did this survey get responses from a representative spread of employees? Or was it all low level employees? Or was it all management and execs? How did the surveyor get a list of employees to survey? Even if the survey came from within Microsoft, its conceivable that they purposely skewed the results to make Steve look bad. Or that they skewed them to make the results as favorable as possible to him. Lot of unknowns.
Ballmer is one of the "goes way back" employees - a friend of Bill who's been there a very long time. If the board decides to replace him with someone who didn't make their bones at Microsoft, it will say a lot. On the one hand, they might need some fresh blood and a new direction. On the other hand, when you bring someone in who didn't grow up in the culture, the company's core values are going to be compromised to a degree.
All the talk of data dedup is great, but in the real world, I wonder if this analogous to another layer of block level compression (I understand the technical difference). A way of saying "These tapes hold (up to) 1600GB", when they really hold 800GB. Some use cases may see dedup as a huge gain, but I don't think we will. My sense is that its a bit gimicky.
I recommended and bought CommVault at version 5.9, and I've stuck with it so far (we've used every version between 5.9 and 8, the current version). It seems to work for the most part, but there are definitely parts that are not polished or streamlined.
With backup software you do a lot of backups and a lot fewer restores. CommVault's backups seem to work. The restores worked for us about 95% of the time. Which is right on the edge of acceptable risk tolerance for us. One restore failed spectacularly and they said "Well, its an issue with the way Windows writes to the NTFS change log." And there was a way to prevent that from being a problem for future backups. But it left me with a lack of confidence that is going to be tough to build on.
You assume that John Young (aka Mr. Cryptome) has one and only one email address. Obviously, he could have (and probably does have) other email addresses.
You assume that there was unencrypted material of value stolen (aside from the 7GB accessible on the Cryptome website.) There's nothing in the article that points to this.
Also, I think it may be more than just a coincidence that Cryptome has 7GB of data, and the hacker said he downloaded 7TB of data. More likely that the hacker typo'd T instead of G. Anybody's who's tried to download a significant fraction of a TB under favorable conditions knows that there are lots of potential problems and things that can go wrong. Its possible that the hacker had been syphoning off data for a long time, but I still think a typo is more likely.
Or you drop it and it breaks?
In short, you're not going to keep much money on your phone. And that will make it easier for someone to switch to a traditional bank or another carrier. And so this market won't have the same barriers to entry that Nokia might wish it had - it should be a fluid market.
I think you're exactly right - the real danger doesn't come from the USA, but from people who would sacrifice one Australian in Sweden to harm the USA.
First off, its not clear where his money is coming from. I would be none too shocked to find out that the CIA were funding him to promote their agenda. Governments make controlled leaks of sensitive information all the time. Maybe wikileaks is one of their pet projects. Maybe it was and he's gone off the reservation.
In any case, Julian Assange is relatively small potatoes in the US. They're not going to "disappear" him or have him suffer a mysterious accident. Too messy and gives too much credibility to wikileaks.
And so the opposite is true too - there are plenty of people who want to know just what will happen if his dead-man's-switch gets flipped. They're the ones he should be worried about.
Maybe there's a conspiracy to get him. Maybe he's part of the conspiracy. Wikileaks isn't the most open organization in the world, you know (see Cryptome's coverage. Short version: Show us the money.)
Or maybe he's a weirdo who has some serious mommy issues.
No, I'm not a lawyer. The one thing he did not do is deny the chargers against him. If I was wrongfully accused of a crime and my lawyer said "Don't say you didn't do it." I'd probably get a new lawyer.
"My cultural values are to protect people's private lives and to never criticize women." (Assange, at around 4:00 in the interview.)
That's a strange way of not denying any of the allegations against him. He's a creep.
For someone who is such a crucader for openness, he has a way of not answering questions directly.
I run my own email - nobody outside of my ISP (easy to switch) or ICANN (not likely) can stop me.
Why would I want to put all that energy into developing for or switching to a service which Google could pull the plug on at their slightest whim?
Google should have worked on disseminating an "OpenWave" reference server - everything you need to run your own Wave. (Of course, storing all of that Wave data is something only Google could do on a large scale, so they'd have little real competition. But people like to know that there are alternatives.)
Once there were some early adopters running their own Wave's, people could potentially have seen the benefits and started looking to adopt Wave. And those people would have flocked to Google.
This is the old Jack Welch (former CEO and Chairman of General Electric) axim - 1, 2, out. For a particular market, you have to have immediate designs on being number one or number two. Third place is not a strategy.
Oracle wasn't going to be number two in HPC any time soon. So they decided to cut their losses. Makes sense if you subscribe to the Welch-ian management style.
According to the Gnome Census, 1% of all contributions to gnome came from Team Ubuntu. (That's 1/16 of what RedHat did, but who's counting...)
See! They give back! They're making the Linux desktop sexy again.
(Of course, they didn't make as many contributions as Nokia, Intel, Fluendo or Novell or of course RedHat.)
"My computer completely stalls when I open my Hotmail. Yes, I have the same issue if I use Firefox, Internet Explorer, or Google Chrome. Yes, I have the same problem if I use another computer, and No, I don't see any script errors because my computer freezes as soon as the window is open... but I have very important emails I have needed to reply to for three days now."
I never thought I would be defending Microsoft, and certainly there are probably cases out there where people are having problems, and I haven't used Hotmail in ages. But...
I don't believe the account of the women quoted. If a web page can lock up a browser or lock up a computer and make it completely unresponsive, there's a flaw in the OS and the browser. And there aren't nearly as many flaws as there once were. And somehow, Microsoft has stumbled onto one that affects three different browsers simultaneously? No.
User error is a more likely explanation - at least in the case above. And you have to question the competency of a user who has emails in Hotmail that are so critical.
I'm not the purist you talk about in your article. I don't compile other people's programs, I like to be as hands off as possible and work at as a high a level of abstraction as possible.
I run a small Windows network and a small Linux network (about 100 machines in total.)
With Puppet on the Linux machines, I can control every aspect of the machine that is necessary - what the configuration files are, which packages are installed, etc. I use the same tool (Puppet) to maintain the state of the machines as I do to get them up and running from a minimal installation. But the key, the real money maker, is that the entire configuration is stored in a version control system. So if I goof up the Kerberos config file, and the users can't login to their workstations... I can just revert back to the previous version, apply the change, and we're done. And of course, the manifests (Puppets answer to GPO's) are in version control too.
With Windows in general, but especially with the Group Policy Management Console, you do the wrong thing, and you have no way of getting an answer to "What changed yesterday?" or "How do I get back to the state where everything was working?" And then, one GPO is easy to test. But when you have 5, things get confusing as to what is being applied and what is being overridden.
I don't necessarily disagree with your assessment here.
But, I think Apple's long term strategy is not to dominate the market (not that they don't want to, but its not a primary goal.) Their goal is to pump out high margin devices. And the margin on Apple products is pretty amazing.
Toyota may sell more cars than Porsche. But that doesn't detract from Porsche.