Follow Slashdot blog updates by subscribing to our blog RSS feed


Forgot your password?

Comment: There are still a lot of Perl shops (Score 2) 269

by oneiros27 (#49746041) Attached to: Ask Slashdot: Career Advice For an Aging Perl Developer?


A couple of years back, I was trying to hire someone ... although we were hoping for OO Perl skills. We ended up hiring someone with database skills to train up in Perl, instead.

The problem with age isn't so much that you have less portable skills, it's that you have a less portable life -- if you have a sponse & kids, you don't want to move the kids in the middle of a school year and away from their friends ... if you have a spouse, you have the problem of trying to find a place that's convenient for both your jobs.

If you're single with no kids ... is hiring in the Netherlands. It's effectively an English speaking country these days (although it's been 30 years since I've been there).

(I have no affiliation with, other than they were a sponsor for many years of the DC-Baltimore Perl Workshop, which I help to organize)

Comment: What is with the hugging? (Score 2) 68

It's not just you -- the article had a quote that makes it seems like there's an unknown sysadmin care bear:

Culturally, we have to make this shift from a mistaken belief that all our data has to be near us and somewhere where I can do and hug the server, instead of someplace where I don't know in the cloud. This is a big shift for many within the department. It's not going to be an easy transition.

I don't know about you, but we don't have our racks just standing their on their own ... they're in rows, so you'd need rather long arms to try to give anything a hug. Maybe the one on the end of the row, but that's still a four or five foot deep rack ... so no hugging if you're short without unracking the gear first.

Oh ... and I've learned to ignore the 'number of data centers consolidated', after it was declared that every wiring closet was a 'data center' in our agency (not DoD), and those were going to count against the 'must close (x%) of data centers'. We 'consolidated' by having two rooms next to each other, so that electrically and AC wise, they're the same room ... but for physical security and access control, it's two separate spaces.

Comment: Suponea their data recorder (Score 1) 615

by oneiros27 (#49707085) Attached to: The Economic Consequences of Self-Driving Trucks

Any vehicle with airbags in it has a data recorder so that the manufacturer can determine under what conditions the air bag did or didn't go off.

So even if you don't have a dash cam, if you're in an accident, get the data from the other vehicle. Depending on the make & model, you might need to make sure that the vehicle doesn't get driven before the data's been read out.

I was once driving home on Thanksgiving day a few years ago -- taking I66 to the DC Beltway. It wasn't rush hour, so I waited until the sign came up that marked that it was now an exit lane (the shoulder is an extra lane during rush hour). It seems that there was some guy *flying* down the lane ... and so he must've considered me to have cut him off.

Once we got to the exit (which had two lanes), he shot around me, then pulled into my lane and slammed on his brakes. When I stopped without hitting him, he gunned his engine, we both got back up to speed ... then he stopped again. I had to stand on my brakes this time.

I kinda wished that I *had* hit him, as I could've proved that it was his fault for wreckless driving. (although a dash cam would've have hurt -- show that he had no reason for braking, as the road was clear, and that he had intentionally gotten in front of me to try to cause an accident).

Comment: and the programmers should not write it (Score 1) 244

by oneiros27 (#49690899) Attached to: RTFM? How To Write a Manual Worth Reading

You can be a programmer and write documentation, but if you want good documentation, the person writing it shouldn't be the same person who wrote the code.

The problem is that you're already too far in -- you understand the design issues, the quirks, etc. You need someone with a fresh view to write the documentation who doesn't come in with a lot of implicit knowledge of the inner workings of the software.

My boss has a policy that it's the newest or youngest person on the project's job to write the documentation, for that very reason. (the others still review it, but they write the first draft).

Comment: Re:FARK has become tamer for various reasons (Score 2) 45

by oneiros27 (#49667545) Attached to: Interviews: Fark Founder Drew Curtis Answers Your Questions

Of course, if you go back to the very early days, Fark didn't have the 'boobies' and 'wieners' topics -- those didn't happen until after they got mentioned in Playboy. (in 2000? 2001?)

(disclaimer : I used to work for Drew, and was the one who added the original topics, back when they all had strong 'S' sounds (and I still thing that 'wierd' and 'cool' weren't needed because of 'strange' and 'spiffy'))

Comment: Not mission creep. (Score 1) 179

by oneiros27 (#49606743) Attached to: NASA Gets Its Marching Orders: Look Up! Look Out!

Could this be an emerging Earth Sciences turf war between NOAA and NASA? Lately it seems more of a National Atmospheric Space Administration. Mission creep, much?

Nope, it's fully in compliance with the 2013 OMB memo on an Open Data policy. The subheading on that memo is 'Managing Information as an Asset', and there is a real lack of a comprehensive catalog of NASA's data. (note that this is *not* the same as the 2013 OSTP memo on public access to federally funded data, but they're related.)

Even with the re-design of, the content behind is is woefully incomplete. When I contacted the creator of the page years ago, he said that they just did some internet searches to find 'data', and then listed them. They were listing websites that mentioned data, not even breaking it down into missions & investigations.

Someone needs to go through and determine for every investigation from every project what data *should* be there, and figure out if it's online, if it's in a dark archive, if the PI still has it, or if it's missing. They should catalog it according to GEMS and possible DataCite (although assignment of 'creator' for the data might be something that needs to be resolved by each science community)

I had tried proposing something to the NASA IT Labs call shortly after the memo came out, but the people running it were blocking our network from being able to submit. I tried again in 2014, and they gave me an alternate way to submit, but they took weeks to get the work-around, and by then I was out of town for a meeting.

(disclaimer : if it's not obvious, I work at a NASA center)

Comment: the joys of other people's closets. (Score 5, Insightful) 106

by oneiros27 (#49436231) Attached to: Bell Labs Fighting To Get More Bandwidth Out of Copper

Back in the 1990s, I was working in Kentucky for an ISP and doing assorted contracting work.

I had a case that was rather similar what you're describing, only ours wasn't run that way because of incompetence -- we were connecting up all of the offices of the Department of Public Advocacy, and for one location the state had decided that rather than get a new line to the DPA offices, as they were in the back of a shopping mall that already had some government offices in it, we'd get fibre pulled between the two offices. Mind you, this was frame relay and fractional T1 days, before DSL, so a new drop was pretty expensive. (I want to say it was around $500/month for just the line charges for a T1, not including the port charges to the ISP tht you were connecting through).

So, when we went there for the install, someone had already pulled the fibre -- I went on the 3 hr drive down there, got soeone to escort me to where I needed to go, and plugged in all of our gear, then went and set things up on the DPA side.

All was fine for a year or so, then we got a call that things were down -- we tried everything that we could over the phone with non-IT folks (it's an office of lawyers), so I was sent on the 6hr round-trip with spare fibre patch cables and such.

A quick check in at the DPA offices showed nothing wrong over there, so I went over the other end of the bulding. I don't remember what the name of the department was, but it was a sort of family services type thing (where people got food stamps, stuff like that). I went up the counter and told the person behind the plexiglass that I was with DPA, and we had equipment in their wiring closet that I needed to get access to.

To which she replied, 'DPA is around the corner'. And I said no, I work for the DPA, and I need to get into your wiring closet. And she kept repeating that DPA was around the corner. I asked for her to get someone else. And I waited 10 minutes or so for someone else to come out front. Once she showed up, I spent a few more minutes with the 'DPA is around the corner' response until I *finally* got through to her and convinced her to let me into their closet. (mind you, this would likely have been considered 'social engineering' if I did it today, as I showed them no ID, being that I had none that said I did work for the DPA).

When I finally got to the closet, I saw that our box had no lights on it ... I traced the power cord down to a power strip that someone had removed all other things from, and taped over those outlets and written 'BAD' across it ... yet left our fiber tranciever plugged into it. I think I was in the room for all of 5 minutes -- it took me *way* more time trying to talk them into letting me in the room than to actually diagnose the problems *including* the time spent in the other offices.

So ~6.5 hrs to fix a problem, because the other office didn't care at all about our gear in their closet, as it would've taken them less than a minute to have moved everthing that was plugged into the known-bad power strip.

So I'd have to say -- no way in hell should you run cable to a private office. If nothing else, that office might close or move, and who knows what might be in there next (or if the new tenents want to remodel it).

Comment: This is new? (Score 1) 121

Admittedly, it wasn't specifically a CompSci class, but when I took our engineering school's 'Intro to Programming' course, we were paired up for the assignments. The only rule was that I wasn't allowed to pair up with Sebastian, as we were the two who had significant programming experience before we got to college.

When I took Numerical Methods my sophomore year, we were paired up in class, but that was partially because the computer lab we worked from didn't have enough computers for all of us. When it came time for the final, they had to book a second lab so that we'd all have computers to compile on. (which meant those of us in the room w/ newer machines had an advantage over the other room, as our code would compile in 1/2 the time)

But let's face it -- group projects are pretty typical in college. And pairing up for labs is normal too ... we don't accuse chemists of getting 1/2 an education if they didn't do every last titration themselves, or a geography major of getting 1/4 an education if they have 3 people in their study group.

The goal is get the people to learn the materials -- if done right, the two people learn from each other. Yes, it can be a drag if you get an idiot for a partner ... but unlike in high school, the people who know their stuff are in demand for their skills, not looked down on for being a nerd/geek/whatever other disparaging term.

If you have two people making forward progress then it's better than one person struggling along and getting nowhere. Maybe I'm being a bit socialist in my views, but there are sometimes when we need to step away from the 'everyone for themselves' typical American attitude and look at the nordic standards for schooling. You don't want your school to get a reputation for being the one that produced someone who screws up in some major way. My undergrad is in civil engineering -- and if I find out I'm in a building that one of my classmates worked on, I'm going to leave ... immediately.

Comment: Reminds me of the spy satellite restrictions (Score 2) 114

by oneiros27 (#49398149) Attached to: DHS Wants Access To License-plate Tracking System, Again

As the government isn't allowed to spy on citizens without a warrent, under normal circumstances, the satellites aren't supposed to take images when over the U.S.

So the government instead buys images from commercial vendors ... the same folks who provide images to Google and Bing for their mapping projects. (which admittedly, might not be as high of resolution).

I'm thinking that there needs to be a line drawn, otherwise all you end up doing is having a way to make an end-run around the legal verdict -- "we'll just spin off a company that does what we're not allowed to do, and buy the results from them".

Comment: Link to the official announcement? (Score 5, Interesting) 122

by oneiros27 (#49352835) Attached to: Amazon Announces Unlimited Cloud Storage Plans

Why do people link to blog posts that neglect to link to the original source?

A little digging, and it seems on the surface to have similar restrictions as BackBlaze, as it's only for "for personal, non-commercial purposes".

So I can't store my ~3PB of telescope data on there, or even just the jpeg browse images.

The terms of use mention that you can share files .. but do they charge you for downloads, as with their other cloud service offerings, or is that included in the 'unlimited'?

  (I might be an old fogey, but I remember when you used to link to a blog post to set context *and* link to the original source in the summary, rather than just some shallow 'I've cherry picked the info'. At least Roland and Coondoggie linked back to their original sources, even if Coondoggies were almost exclusively regurgitation of press releases + a links back to Network World))

Comment: Useful in what way? (Score 2) 47

by oneiros27 (#49348091) Attached to: Is the Apple Watch a Useful Medical Device? (Video)

I've seen a few talks from Stephen Friend. I was at the Research Data Alliance meeting, and he gave one of the plenary talks the day after Apple unveiled the device, and announced Research Kit (which he's involved with).

He mentioned that less than 24hrs after its release, they already had more Parkinsons patients signed up than any published study on the disease.

If the watch can get *any* sort of medically useful data, I'm all for it, especially as so many people have been designating that their data can be used by any qualified researcher. (yes, there will still have to be IRBs to approve research at most institutions, and I assume some sort of gatekeepers from Sage Bionetworks to determine who gets access to the data). ... but the fact that we might be able to get medical data at a scale never before seen is huge. And we might get a wider slice of the population, not just college students or from a limited geographic area that might not be applicable to the larger population.

(disclaimer : I did not watch the video. I usually read the articles before commenting (I know, that's against this site's standards) ... if the person has a legitimate argument to make, post it so I can read it)

Comment: who? people who got stung by ExtJS (Score 1) 320

by oneiros27 (#49306447) Attached to: Why I Choose PostgreSQL Over MySQL/MariaDB

When ExtJS changed their license to GPL3, not LGPL, as you would expect for a library.

The owner of Sencha then put out a statement that if you built something that made use of ExtJS, then you had to release your software under GPL3 ... including the server components.

I have no problem with releasing the client side -- that's all javascript that people could view the source and see ... but releasing the server side? That requires security audits and a review by legal ... it's just not going to happen.

Reading the review, the reviewer seemed to have the same take on what GPL meant from the statement :

With MySQL, on the other hand, the client library is GPL, so you must pay a commercial fee to Oracle or supply the source code of your application. (Thatâ(TM)s less of an issue when using MySQL in websites; MariaDB uses the GPL 2 license but also has a less restrictive LGPL license for MySQL Client libraries.)

Now, if the issue is simply the *client* code, then you could get around it by using ODBC, or something like Perl's DBD::mysqlPP, which doesn't use the MySQL client code. Do you have to release the whole application if it's just something that makes use of a mysql database? I don't know, but with all other things being equal, and more and more people coming to this conclusion, I'd rather just stick with something that's LGPL or MIT.

Comment: Re:No excuse? BS. (Score 1) 155

by oneiros27 (#49283805) Attached to: White House Proposal Urges All Federal Websites To Adopt HTTPS

Who's going to pay for the CDN? My data is growing at > 1TB/day, and I have no idea what's going to be of interest on any given day.

And as for CPU cost ... are you going to pay for the sysadmin time to migrate all of our services? Or any of the other solutions that you're proposing?

Our servers have been certified as 'low' risk for years, because we're specifically distributing data with *no* access restrictions. We've had to fight for our 'low' ... and then have to explain to the security auditors every three years that what they're testing for doesn't apply to us.

(we have one of the highest 'incident' rate for our location, because they consider every attempt at a hack to be a 'incident', even though we haven't had any successful hacks in years).

Oh ... and of our staff of 2.5 sysadmins for our department, dealing with security audits and such takes up > 0.5 FTE for about 6-9 months or so when the security plans are updated and the audits are occuring ... so it's not cheap).

No more unfunded mandates ... if this is important enough ... give us the funding and resources to do it. (which likely means hiring another sysadmin, and more hardware)

I'd go back to FTP before I went to HTTPS.

Comment: No excuse? BS. (Score 2) 155

by oneiros27 (#49282307) Attached to: White House Proposal Urges All Federal Websites To Adopt HTTPS

I operate government websites that serve physics data to the public.

HTTPS would require additional CPU for the SSL processing and bandwidth because it would make requests non-cacheable.

Not to mention that it would make the intrusion detection system attached to the router completely useless, so we'd lose a layer of security and it would make it more difficult to detect probing across the network and other 'slow' attacks. It would also prevent us from doing auditing after an exploit is known but before we've been able to get the mod_security rules in place or whatever other mitigation.

So yes, there are perfectly valid reasons to *not* be running HTTPs. I know you couched your message with 'virtually', but blindly appying 'best practices' or whatever other recommendations without understanding what the implications will break systems. (and I have to file paperwork every year for every one of my web servers that doesn't comply with the CIS benchmarks)

ps. 'there should be a law for that' is the absolutely worse policy, as most people in legislature aren't tech-savy, and will just screw things up. I was actually against all of the Net Neutrality bills that were proposed because they'd have outlawed agressive spam filtering (blocking 'legal' communications, and the CAN-SPAM act defined that some spam is legal). You need flexibility and speed in dealing with most issues, and laws don't do either well.

Comment: Conflicts w/ his first biography (Score 1) 205

by oneiros27 (#49258367) Attached to: Steve Jobs's Big Miss: TV

At least, I was assume it was in his biography (as I never read it). But when it came out, there were quite a few reports that Jobs said he had figured out TV interfaces:

It's entirely possible that because he didn't like the TVs, he had come up with a better UI ... but we haven't seen a dramatic revision of the Apple TV since he died ... so we might never know what it was that he came up with.

Top Ten Things Overheard At The ANSI C Draft Committee Meetings: (9) Dammit, little-endian systems *are* more consistent!