Computing Blog

A blog about all aspects of computing and technology from software development to social network to commentary on the IT industry as a whole.

Rokoâ€™s Basilisk â€“ Lovecraftian Calvinism on Steroids?

Posted on July 19, 2014 by Tim Hall

Memetic_hazard_warning Slate Magazine has discovered Rokoâ€™s Basilisk: The most terrifying thought experiment of all time, which postulates that an all-powerful Godlike artificial intelligence will punish everyone who didnâ€™t help it come into existence in a computer-generated afterlife.

SF author Charlie Stross blogged about Rokoâ€™s Basilisk last year, and correctly identified is an a nasty mashup of the bleakest elements of Calvinist theology with H.P.Lovecraftâ€™s â€œThings Man Was Not Meant To Knowâ€.

Leaving aside the essentially Calvinist nature of Extropian techno-theology exposed herein (thou canst be punished in the afterlife for not devoting thine every waking moment to fighting for God, thou miserable slacking sinner), it amuses me that these folks actually presume that weâ€™d cop the blame for itâ€”much less that they seem to be in a tizzy over the mere idea that spreading this meme could be tantamount to a crime against humanity (because it DOOMS EVERYONE who is aware of it).

And now I discover Iâ€™m followed by Rokoâ€™s Basilisk on Twitter. Should I be worried?

Posted in Computing, Religion and Politics, Science Fiction | Tagged Artificial Intelligence, Charlie Stross, Moonbats | Comments Off

Should you log every bug?

Posted on June 18, 2014 by Tim Hall

Are all bugs serious enough to log?

I found one during testing today that raised the question of the costs of raising, triaging, fixing and retesting a bug compared with the costs of just leaving it in the system.

It set off some interesting discussions on Twitter on the subject.

@kinofrost @Kalyr Yes. If logging a bug is more likely to add more cost than value, don't log it — for now. … but do mention it.

— Ben Simo (@QualityFrog) June 18, 2014

@Kalyr Are you measured by how many bugs you find? More serious answer – are you sure you are able to assess the triviality of it?

— pkirkham (@pkirkham) June 18, 2014

@QualityFrog @kinofrost @Kalyr If it bugs you then record else not?If you are only person in the forest hearing thunder did someone else did

— Mohinder Khosla (@mpkhosla) June 18, 2014

@mpkhosla @kinofrost @Kalyr Not recording doesn't mean I tell no one. Trivial bug may be a costly distraction, or may not be trivial.

— Ben Simo (@QualityFrog) June 18, 2014

@kinofrost To my experience "baby dragons" like that can grow big. Could be good to MIP or log somewhere. @QualityFrog @mpkhosla @Kalyr

— Annie Rydholm (@AnnieRydholm) June 18, 2014

The bug in question was a drop-down list for an optional value in a create dialog in a relatively little-used part of the system. One of the values in the dropdown was not valid, and quite obviously not so, at least with the data in the test system. If you selected that bogus value and saved the record, the server-side code didn’t throw an error, but inserted null into the column.

Since neither the field nore the underlying column were mandatory, the bug didn’t result in any data corruption, and it did save the correct value to the database if you selected one of the legitimate, valid values.

From a user perspective, it’s more of an irritation than a major issue; it certainly doesn’t prevent you from using the feature in its intended way.

In the end I discussed it with the business analyst, and we agreed to log it in the bug tracking system, but assign a low severity.

Posted in Testing & Software | Tagged Bugs, Twitter | 2 Comments

Is This A Bug?

Posted on June 17, 2014 by Tim Hall

On the reset password function, Username is case-insensitive.

On the login function, Username is case-sensitive.

So if you try to log in with the exact same username as you used to reset the password (i.e, with the “wrong” capitalisation”), and the brand new password you just created, it’s rejected as “Invalid username and password”.

Logon and password reset features are basic functionality for all websites. But still some people get it wrong…

Posted in Testing & Software | 2 Comments

Will Everything Always Be Broken?

Posted on June 8, 2014 by Tim Hall

This piece entitled “Everything Is Broken” describes the dire state of internet security, with major security breaches every other day, with everyone from Russian criminals to The NSA making off with our data.

It was my exasperated acknowledgement that looking for good software to count on has been a losing battle. Written by people with either no time or no money, most software gets shipped the moment it works well enough to let someone go home and see their family. What we get is mostly terrible.

It paints a very depressing picture of an internet held together with sellotape and string, but do you know what it reminds me of?Â Replace catastrophic data breaches with fatal accidents, and it reminds me of the railway industry when it had been around for about the same length of time as the internet has, somewhere in the middle of the 19th century.

If you read L.T.C.Rolt’s classic “Red For Danger”, the mid-Victorian railways suffered serious crashes on a regular basis. Primitive signalling systems were vulnerable to human error. Braking systems were crude and ineffective. And flimsy wooden carriages with gas lighting were reduced to matchwood in relatively low-speed collisions and often went up in flames.

But things got better. It took many years, but eventually a combination of legislation and market pressure saw safety taking a much higher priority, and serious crashes are now few and far between.

It’s anyone’s guess what the internet will look like in a century’s time. But it’s entirely possible that netizens of the 22nd century will look back at the data breaches and insecurity of today like we look upon 19th century industry.

Posted in Testing & Software, Uncategorized | Tagged Security | Comments Off

Always Back Up Your Data

Posted on June 3, 2014 by Tim Hall

It Bites have had to cancel their appearance at Prog Splash in Holmfirth because they’ve lost all their data:

It Bites have been forced to pull out of Saturday’s Holmfirth Prog Splash. Due to technical issues and a memory card malfunction John Beck’s entire It Bites catalogue of keyboard programming – some 30 years worth – has been lost.

Under normal circumstances in the time between now and the gig John would have worked day and night to reprogram it but with him committed to other projects this week the task has proved impossible. Therefore, rather than put on a sub standard show the band have taken the difficult decision to cancel the gig.

It Bites realise this impacts on both their fans and to other bands on the bill and sincerely apologise to all.

The moral to this is “always back up your data”. Or go back to the good old days of Hammond B3s and Mellotrons….

Posted in Music News, Testing & Software | Tagged It Bites | 3 Comments

Is this a Bug or an Issue?

Posted on May 21, 2014 by Tim Hall

SNCF Wide Train The French railways seem have have found a serious bug in integration testing. As reported in BBC news, French red faces over trains that are ‘too wide’

The error seems to have happened because the national rail operator RFF gave the wrong dimensions to train company SNCF.

Our correspondent says that they measured platforms built less than 30 years ago, overlooking the fact that many of France’s regional platforms were built more than 50 years ago when trains were a little slimmer.

This is a prime example of a bug which would have been an awful lot cheaper to fix had it been caught at the design phase of the project.

Posted in Testing & Software, Travel & Transport | Tagged SNCF | 4 Comments

Things Twitter could do

Posted on May 16, 2014 by Tim Hall

I think few people would deny that Twitter has a troll problem. For us regular users with a few hundred followers it’s easy enough to block the occasional drive-by troll, especially if we’re male. But it’s a different story for public figures, especially women, who can find themselves bombarded with hundreds of abusive messages.

Technical solutions for social problems aren’t ideal, but trying to re-educate the sections of the population who live in the bottom half of the internet is at best a very long term project.Â In the meantime there are things Twitter could do make it harder for trolls to ruin people’s Twitter experience.

One would be to give users the ability to filter the Notifictations tab. At the moment, anyone you haven’t blocked will be visible in that tab if they @message your username. It’s not technically difficult to filter than by degrees of separation, so what you see in your Connections tab can take into account things like:

The number of people you’re following who follow them
The number of people you follow who have blocked them
The total number of people who have blocked them relative to their number of followers.

Of course it would need to be refined to prevent the trolls themselves from gaming the system. For example, perhaps blocks from those who are very block-happy but have themselves collected a lot of blocks could be disregarded.

Twitter could also crack down on abuse of multiple accounts. There are plenty of legitimate reasons why people need multiple accounts, but it’s well known that trolls often churn through multiple throwaway accounts as each one gets blocked by their targets. Surely it’s not impossible for some kind of pattern-matching on IP addresses and word use to identify which accounts are being used by the same, and deal with them accordingly when any one is suspeded for abuse.

Twitter is very efficient at nuking spam accounts, and they’re pretty easy to identify algorithmically. Dealing with trolls is harder, and will require more human intervention, but that’s no excuse for Twitter to do nothing. As I’ve pointed out, there are plemty of things they could do if the will was there.

Posted in Social Media | Comments Off