Posts tagged ‘Amazon’


Why you shouldn’t build projects on Google Cloud (or, why are we still having these conversations?)

01.07.2018

This story on Medium, about Google Cloud, is all too familiar to me (hat tip to Donkey). It mirrors my experiences with Google in 2009 and 2013.
   A company monitoring solar plants and wind turbines had Google pull their account twice. The Googlebot falsely claimed there was suspicious activity, with Google threatening to delete their account in three days. If their CFO, whose credit card was linked to the account, hadn’t replied in time, that would have been millions of dollars down the gurgler.
   The company’s warning: don’t build stuff on Google Cloud. Apparently AWS is safer.
   They were very lucky, because Google’s forums are littered with people whose accounts have also been unilaterally terminated, and they were never recovered. Some have lost income streams. Most went through the “proper channels”.
   My experience in helping a friend recover his blog nearly a decade ago showed just how unreliable these channels were, with a Google forum volunteer going out of his way to be obstructive, because you dared question the big G. Most volunteers actually seem offended you questioned Google, such is their adherence to the cult.
   Mind you, I’m still waiting, three years later, for an explanation about why our Amazon Associates’ account, nearly two decades old, was unilaterally deleted. Amazon claimed six months ago that the matter had been ‘escalated’. Still waiting. Google, too, gets back to you initially, but escalation results in nought.
   When things go wrong, US Big Tech doesn’t work, does it? We’ve actively avoided Google for nearly a decade, and began posting warnings about Facebook around that time, too.
   Thank goodness for companies like Zoho: in the 2010s, Indian tech works better, and people take greater responsibility.

Tags: , , , , , , ,
Posted in business, internet, technology, USA | No Comments »


The path of least resistance: we humans aren’t discerning enough sometimes

04.02.2018

I came across a thread at Tedium where Christopher Marlow mentions Pandora Mail as an email client that took Eudora as a starting-point, and moved the game forward (e.g. building in Unicode support).
   As some of you know, I’ve been searching for an email client to use instead of Eudora (here’s something I wrote six years ago, almost to the day), but worked with the demands of the 2010s. I had feared that Eudora would be totally obsolete by now, in 2018, but for the most part it’s held up; I remember having to upgrade in 2008 from a 1999 version and wondering if I only had about nine years with the new one. Fortunately, it’s survived longer than that.
   Brana Bujenović’s Pandora Mail easily imported everything from Eudora, including the labels I had for the tables of contents, and the personalities I had, but it’s not 100 per cent perfect, e.g. I can’t resize type in my signature file. However, finally I’ve found an email client that does one thing that no other client does: I can resize the inbox and outbox to my liking, and have them next to each other. In the mid-1990s, this was one of Eudora’s default layouts, and it amazed me that this very efficient way of displaying emails never caught on. I was also heartened to learn from Tedium that Eudora was Apple co-founder Steve Wozniak’s email client of choice (‘The most important thing I use is Eudora, and that’s discontinued’). I’m in good company.
   However, this got me thinking how most users tolerate things, without regard, in my opinion, to what’s best for them. It’s the path of least resistance, except going down this path makes life harder for them.
   The three-panel layout is de rigueur for email clients today—all the ones I’ve downloaded and even paid good money for have followed this. Thunderbird, Mailbird, the oddly capitalized eM. All have had wonderful reviews and praise, but none allow you to configure the in- and outbox sizes. Hiri’s CEO says that’s something they’re looking at but right now, they’re not there, either. Twenty-plus years since I began using Eudora and no one has thought of doing this, and putting the power of customization with the user.
   But when did this three-panel layout become the standard? I can trace this back to Outlook Express, bundled with Windows in the late 1990s, and, if I’m not mistaken, with Macs as well. I remember working with Macs and Outlook was standard. I found the layout limiting because you could only see a few emails in the table of contents at any given time, and I usually have hundreds of messages come in. I didn’t want to scroll, and in the pre-mouse-wheel environment of the 1990s, neither would you. Yet most people put up with this, and everyone seems to have followed Outlook Express’s layout since. It’s a standard, but only one foisted on people who couldn’t be bothered thinking about their real requirements. It wasn’t efficient, but it was free (or, I should say, the licence fee was included in the purchase of the OS or the computer).
   ‘It was free’ is also the reason Microsoft Word overtook WordPerfect as the standard word processor of the 1990s, and rivals that followed, such as Libre Office and Open Office, had to make sure that they included Word converters. I could never understand Word and again, my (basic) needs were simple. I wanted a word processor where the fonts and margins would stay as they were set till I told it otherwise. Word could never handle that, and, from what I can tell, still can’t. Yet people tolerated Word’s quirks, its random decisions to change font and margins on you. I shudder to think how many hours were wasted on people editing their documents—Word can’t even handle columns very easily (the trick was usually to type things in a single column, then reformat—so much for a WYSIWYG environment then). I remember using WordPerfect as a layout programme, using its Reveal Codes feature—it was that powerful, even in DOS. Footnoting remains a breeze with WordPerfect. But Word overtook WordPerfect, which went from number one to a tiny, niche player, supported by a few diehards like myself who care about ease of use and efficiency. Computers, to me, are tools that should be practical, and of course the UI should look good, because that aids practicality. Neither Outlook nor Word are efficient. On a similar note I always found Quattro Pro superior to Excel.
   With Mac OS X going to 64-bit programs and ending support for 32-bit there isn’t much choice out there; I’ve encountered Mac Eudora users who are running out of options; and WordPerfect hasn’t been updated for Mac users for years. To a large degree this answers why the Windows environment remains my choice for office work, with Mac and Linux supporting OSs. Someone who comes up with a Unicode-supporting word processor that has the ease of use of WordPerfect could be on to something.
   Then you begin thinking what else we put up with. I find people readily forget or forgive the bugs on Facebook, for example. I remember one Twitter conversation where a netizen claimed I encountered more Facebook bugs than anyone else. I highly doubt that, because her statement is down to short or unreliable memories. I seem to recall she claimed she had never experienced an outage—when in fact everyone on the planet did, and it was widely reported in the media at the time. My regular complaints about Facebook are to do with how the website fails to get the basics right after so many years. Few, I’m willing to bet, will remember that no one’s wall updated on January 1, 2012 if you lived east of the US Pacific time zone, because the staff at Facebook hadn’t figured out that different time zones existed. So we already know people put up with websites commonly that fail them; and we also know that privacy invasions don’t concern hundreds of millions, maybe even thousands of millions, of people, and the default settings are “good enough”.
   Keyboards wider than 40 cm are bad for you as you reach unnecessarily far for the mouse, yet most people tolerate 46 cm unless they’re using their laptops. Does this also explain the prevalence of Toyota Camrys, which one friend suggested was the car you bought if you wanted to ‘tell everyone you had given up on life’? It probably does explain the prevalence of automatic-transmission vehicles out there: when I polled my friends, the automatic–manual divide was 50–50, with many in the manual camp saying, ‘But I own an automatic, because I had no choice.’ If I didn’t have the luxury of a “spare car”, then I may well have wound up with something less than satisfactory—but I wasn’t going to part with tens of thousands of dollars and be pissed off each time I got behind the wheel. We don’t demand, or we don’t make our voices heard, so we get what vendors decide we want.
   Equally, you can ask why many media buyers always buy with the same magazines, not because it did their clients any good, but because they were safe bets that wouldn’t get them into trouble with conservative bosses. Maybe the path of least resistance might also explain why in many democracies, we wind up with two main parties that attract the most voters—spurred by convention which even some media buy into. (This also plays into mayoral elections!)
   Often we have ourselves to blame when we put up with inferior products, because we haven’t demanded anything better, or we don’t know anything better exists, or simply told people what we’d be happiest with. Or that the search for that product costs us in time and effort. Pandora has had, as far as I can fathom, no press coverage (partly, Brana tells me, by design, as they don’t want to deal with the traffic just yet; it’s understandable since there are hosting costs involved, and he’d have to pay for it should it get very popular).
   About the only place where we have been discerning seems to be television consumption. So many people subscribe to cable, satellite, Amazon Prime, or Netflix, and in so doing, support some excellent programming. Perhaps that is ultimately our priority as a species. We’re happy to be entertained—and that explains those of us who invest time in social networking, too. Anything for that hit of positivity, or that escapism as we let our minds drift.

Tags: , , , , , , , , , , , , , , , , , , , , , , , , , , , ,
Posted in business, cars, culture, design, internet, politics, publishing, technology, USA | 2 Comments »


It took a little longer, but Autocade reaches 12 million views

03.01.2018

It’s a little disappointing to note that Autocade has taken slightly longer to reach 12 million page views: it ticked over to its new milestone earlier today. I really had hoped that we’d get there before 2017 was out, but it was not to be.
   Part of it might have been the slower rate of models being put up—life’s been busy, and a site that earns a fairly small amount of money compared to our other businesses doesn’t warrant as much time. But 100 models have gone up since June 2017, when Autocade reached its 11 million milestone, with the 3,600th model the Nissan Rasheen (and no, I didn’t plan this one—it’s quite an oddball vehicle).
   So here’s the running tally as I’ve been keeping on this blog, for really no reason other than pedantry.

March 2008: launch
April 2011: 1,000,000 (three years for first million)
March 2012: 2,000,000 (11 months for second million)
May 2013: 3,000,000 (14 months for third million)
January 2014: 4,000,000 (eight months for fourth million)
September 2014: 5,000,000 (eight months for fifth million)
May 2015: 6,000,000 (eight months for sixth million)
October 2015: 7,000,000 (five months for seventh million)
March 2016: 8,000,000 (five months for eighth million)
August 2016: 9,000,000 (five months for ninth million)
February 2017: 10,000,000 (six months for tenth million)
June 2017: 11,000,000 (four months for eleventh million)
January 2018: 12,000,000 (seven months for twelfth million)

   It’s a shame that the four-month time-frame needed to reach 11 million could be an anomaly rather than part of a trend.
   I also wonder whether the odd PHP error—we have had quite a few since we began hosting at AWS—has impacted on search-engine rankings. However, server management has become far, far more complex over the last couple of decades, and the controls I see at AWS mean nothing to me as someone outside the computing industry. The help pages may as well be in Serbian. The notion that software gets easier to use and the expectation that this level of computing would become democratized have not come to pass, certainly not over the last 10 years. It seems the industry wants to sew things up for itself, and the last thing needed are amateurs like me getting into the nuts and bolts. I’m not Facebook or Google: I can’t afford heaps of employees to look after this stuff. (Or, in Google’s case, maybe a