varunnangia.com

The picture is of two identical 60W MagSage Power Adapters, the kind that came stock with 2009 “Unibody” Apple MacBooks that year. They came with machines bought 42 days apart in 2009. BAGKA, the one on the left, was used with my machine which was bought first, and saw a lot of use and is literally scuffed from use. YAGKA, on the right, came with the second MacBook 42 days later, and the machine saw next to no use and literally sat at home plugged in carefully. YAGKA looks new, smells new, feels new – the power cord still has that slightly rough sandpaper feel that all new MagSafe adapters cords have.

42 days ago, BAGKA failed without warning. In the morning, it was fine. In the evening, it didn’t charge either computer. If you stuck your ear next to it, you could hear a steady static buzz that didn’t change whether it was plugged in or not. The next day, I went out and bought a new power adapter because I thought I’d destroyed BAGKA with use.

Today, YAGKA failed without warning. In the morning, it was fine. In the evening, it stopped charging, and it’s making the same static buzz noise that BAGKA did.

So my question is: as unnerving as it is to have these fail exactly the same number of days apart as we bought them, did we get unlucky? Or, are we caught up in some wider quality problems that Apple hasn’t spoken up about yet?

With any luck – *knocks on wood* – my current laptop, a 2009 MacBook, will last me another 2-3 years. I’ve maxed out the memory, replaced the primary drive with an SSD and removed the optical drive in favour of a second mass storage hard drive. I’m still on 10.6.8, rather than 10.8, which would enable the possibility of creating a “Fusion Drive“, but now more because of “it works, don’t break it” rather than the compelling reason I had earlier – a 32bit kext for which no 64bit kext will ever be made.

However, in the wake of my third failed attempt to replace the prematurely dead internal battery, I got to thinking about the next laptop I would buy. For many reasons, I am reluctant to buy another Mac. The one reason I would is because of the Magsafe connector, which means that when the cats go flying past, the laptop doesn’t go with them. It’s literally been a lifesaver for the machine; it’s also been incredibly convenient. Since the USPTO has, in its infinite (lack of?) wisdom awarded Apple a patent for an obvious connector, there are really no other machines available with this incredibly useful feature – possible exception, the Surface and I’m not sure if it’s a pogopin design.

Meanwhile, I have not seen anything compelling in the latest Mac OS releases – Fusion Drive possibly excepted. The iOSification of Mac OS means that increasingly Mac OS doesn’t offer me anything that cannot be done elsewhere. At the same time, I have moved the rest of my computing life to Android, which is increasingly more capable, and quickly replicating the power and flexibility that I used to think was possible only on a desktop OS. If indeed Android eventually gains windowing support, I could actually see myself using that full time.

Fortunately this is still a hypothetical discussion, but I’m hoping there will be more choices by the time this machine eventually needs to be consigned to history.

I took a few minutes to start playing Halo 4 today and I was again reminded fairly early on in the game why it is I have always loved the series: the colour palette. Unlike other games that use predominantly one portion of the spectrum – think of the reds or oranges of Gears of War or dark browns and grays of Call of Duty – Halo has always had dark indoor scenes interspersed with bright blues and greens of a perfect sunlit day outdoors.

To this day, I rue the closure of the Halouvre, a gallery of some of this visual wonders – now you’ll either have to play the game, or take our word on it.

Or as we would say on the internets: JUST LOOK AT IT. And try not to die in the process of gawking.

It’s 2012, flash costs ≈$0.25/GB and I have to manage every KB carefully because I can’t buy a stock Android device with more than 16GB* of storage space. #fail

*: <13GB usable.

PS: Pic shows result after removing apps and music of about 1.25GB.

Well, shit. Pray this disk isn’t going bad; this isn’t easy to replace.

I was dragged to see Ridley Scott’s Prometheus, largely against my will. Now I quite like Scott, and both Kingdom.of Heaven and Bladerunner are amongst my favourite movies, but this was … terrible (and this might be a generous assessment).

To be sure, the space porn was spectacular, with some truly stunning visual effects of the ship itself or it’s arrival at the target planet. But from thereon, it fell apart, largely on its slavish following of every terrible sci-fi cliche in existence. Consider:
• Bad robot that knows more about the mission than the humans? Check.
• Willy nilly headlong dive into uncharted territory with no assessment of risk? Check.
• Greedy rich old corporation with hidden agenda funding the mission? Check.
• Token minority dude spouting one-liners? Check, again.

In fact, the only thing different about this film from, say Transformers or , is the use a flamethrower as the primary weapon, instead of some laser-based weapon. Meanwhile, as I read Kin Stanley Robinson’s 2312
, I again curse the lack of an imaginative, unusual, boundary-pushing sci-fi story being taken from word onto screen.

And last a word to people classifying this as a sci-fi story: it is not. Prometheus, like Sunshine from a few years ago, is primarily a horror flick (something I loathe) set against a science fiction background, but setting it in the present or past would work as well. To call it a sci-fi movie would be like calling the movie versions of Another Earth or The Time Traveler’s Wife sci-fi, instead of a romance film. A futuristic setting doesn’t make for sci-fi: something pushing our idea of what is normal means coupled with something that inspires people to build that future is sci-fi; the rest is just noise, like Prometheus.

Shame. I had high hopes.

Via Katy Hospital, this depressing infographic about the US’ slide into pre-Enlightenment morass. Seriously considering moving states – collective human rights outweigh my economic rights, any day.

Remember: conservatives worship dead liberals.

My earlier post indicated that there were some issues running an Agility 3 within a Macbook, but what I hadn’t realized was that there were two sets of problems that I was seeing:

1. Issues that are general to all Macs using non-Apple SSDs; and,
2. Issues that are specific to the make/model of the Macbook (MacBook 6,1) I have and the SSD (Agility 3) I have.

In fact, the only issue that falls into the first category is the lack of native TRIM support, and that is fixable with the TRIM Enabler tool. Everything else, unfortunately, falls into the second bag of hurt, so here’s what you need to know.

There is a known issue with the NVIDIA GeForce 9400M chipset used in many 2009-era Mac Mini, MacBooks and MacBook Pros and the Sandforce-2281 controller. Specifically, the 9400M and the SSD fail to negotiate the link speed correctly, and every time the drive comes out of idle, it manifests as a stutter or beachball. Therefore all of my earlier suggestions worked primarily because they stopped the drive from going into idle as often as it normally would, but did nothing for the underlying issue; I treated the symptom, not the disease.

Is your negotiated link speed showing as 3 Gigabit?

To confirm that this is what you’re seeing, fire up System Profiler, navigate to the Serial-ATA section, and see whether the “Negotiated Link Speed” is showing as anything other than 3 Gigabit. If it’s showing as 1.5 Gigabit, then you’re suffering from this issue. The image at the right shows you what you’re looking for.

The correct fix for this is to go to OCZ’s forum and download the Bootable Tools CD, and burn it to a disc. Restart the machine, holding down the “C” key to boot from the optical drive. Once booted into the Tools, click the icon at the bottom that has two green arrows pointing to the right and hit Enter Return when prompted. This permanently locks the drive at 3 Gigabit mode, instead of permitting the drive to negotiate the speed. As long as you’re using the drive with a machine that uses a NVIDIA GeForce 9400M chipset, this is fine – you can’t go above 3 Gigabit and it won’t fall back to 1.5 Gigabit if it can’t negotiate with the drive and it’ll prevent the constant renegotiation that causes beachballing and stuttering.

Once done, go ahead and use TRIM Enabler to enable TRIM and restart. You should be good to go. Enjoy the future.

(PS – if you take the SSD out and put it in a non-9400M machine, follow the same instructions to start off the Bootable Tools CD, and then click the icon with the two red arrows pointing to the left. Done.)

Update: this is wrong and was based on an incorrect understanding of what was happening; go here for instructions to apply the correct fix (and a brief explanation as to what’s happening and why).

SSDs have finally come down in price to about $1/GB – very pricey still compared to spinning platters – but reasonable compared to some of the alternatives. So when a combination of rebates, sales, and gift cards aligned, I picked up an SSD a few weeks ago and put it into my Macbook that’s running Mac OS X 10.6 “Snow Leopard” to see if it cleared up performance issues I was having.

And the answer is very emphatically yes. I can literally restart in the time it took me to launch Firefox with the old spinning drive.

However, it’s not been all roses – neither Sandforce nor OCZ Technology, the controller maker and the drive maker, respectively, support Mac OS in anything other than name, so fixing issues has been something of a challenge. I would say, based on my experience, that if you want an SSD in your Mac, and you are not technically inclined, pay the Apple premium and get the official SSD.

First issue: I’ve been vexed by a “stutter” or “freeze” problem since I installed the Agility 3 in here. As of today, the problem doesn’t appear to occur anywhere except in the first few seconds of Chrome or Safari, neither of which I use very often, so I’m marking the issue as closed.

The other issue that I think is resolved occurs at startup. From the time the blue background kicks in to when my desktop has finished loading is somewhere in the vicinity of 7-8seconds. However, I somehow remember the bit before – the gray screen section – as being faster on a spinning drive. This is perception-based though; it may actually be as fast, but because the blue Mac OS X loading section was so much longer, this seemed faster.

So a few tips based on my experience.

1. Do not use the TRIM Enabler tools out there. Just turning this off reduced the stutter by about 80 percent – I suspect it’s because the driver is too aggressive when it comes to issuing TRIM commands to the drive. I’ll keep my eye on the development to see if it improves, but for now, I’m happier with it off. (Of course, this is a Sandforce drive, so don’t back it into the corner of doom – keep about 15 percent free, on top of what the drive firmware reserves).
2. Verify and repair the disk and disk permissions. Twice each. My clone was apparently not perfect, despite CCC’s assurance everything was copied over perfectly. This halved the remaining stutter and stopped me finding “Recovered files” in the Trash every time I booted.
3. Reset the SMC and then reset the PRAM, in that order. This reduced the apparent time I spend on the gray screen by about a third, but the first boot after wiping the PRAM was very time consuming.
4. In the “Energy Saver” item of System Preferences, uncheck the options to allow the drive to sleep when it is idling. This is the bit I figured out today. The stutter seems to have vanished once I did this, Chrome and Safari excepted.

It’s not all roses, but it’s getting there – I love living in the future.

Though some might say “tl;dr”, this article caught my eye not just as a review of an interesting piece of hardware (which it is), but because of what is happening in the generic ARM space.

We’re not quite where we are with easily swappable and standardized x86/x64-powered hardware, but we’re getting there. These designs remind me of early x86 hardware, which were commodity parts, but also didn’t have standardized form factors, port locations, RAM setups, BIOSes and so on.

I’m guessing we’re all of 2-3 years away from being able to get a standard mATX or ITX ARM-powered motherboard, which we’ll be able to boot anything on, attach the usual battery of hardware to and so on. Possibly, we’ll be able to upgrade the chips by swapping them out.

Glad to see Phoronix is ahead of the game by getting ready for that future with its benchmarks…