Karabiner used to be a solution, with a feature that required ⌘Q to be pressed twice before the app would quit , but the Karabiner Elements, the MacOS Sierra version, doesn’t support this.

However, there is a fairly simple(ish) way to recreate this behaviour, using Hammerspoon, a ” a desktop automation tool for OS X. It bridges various system level APIs into a Lua scripting engine”.

So, to prevent cmd-q from instantly quitting the current app do the following:

• install hammerspoon (http://www.hammerspoon.org/go/)
  • download latest version
  • drag into /Applications folder
• Run Hammerspoon
  • Hammerspoon saves its config in $/.hammerspoon, and we can add our own scripts here
  • The Lua script to prevent Cmd-Q killing an app is here
  • Save the lua script as ~/.hammerspoon/double_cmdq_to_quit.lua
  • Edit ~/.hammerspoon/init.lua and add this line: require("double_cmdq_to_quit")
  • Reload config from Hammerspoon’s menu bar icon

Typing ⌘Q should now show this

and typing ⌘Q will now quit whichever app we were trying to quit.

( thanks to the good folk in this GitHub issue https://github.com/tekezo/Karabiner-Elements/issues/242)

The post Fixing CMD-Q annoyances appeared first on digitalquery.

However, nobody was fooled by that and nearly a year later, I’m here to try again. Yes, the site’s been redesigned again,  once again taking its design cues from the previous year’s (or the year before that) hot design trends. I look forward to once again spending the next twelve months watching it date horribly with every day that passes.

However, on a brighter note, i’m very happy with how it’s been built. It was thrown together in a weekend, using a new theme foundation we’ve been working on, and which attempts to add an  MVC layer on top of WordPress. Integrating in a set of amazing plugins, it will allow us to deliver very feature rich sites more quickly, and also allow the client a huge amount of control over the look and feel of their pages.

I’m also reasonably content with our tooling – we’ve got a nice Vagrant based development environment set up, based on WP Vagrant, a Vagrant configuration for WordPress that I put together last year. We’ve settled on Zurb Foundation for our front end needs, and PhpStorm as the IDE. BitBucket or GitLab as our remote git server, and Capistrano for deployment.

Also, beefing up the infrastructure, last month, I set up a 32GB bare metal server from Hetzner as an OpenVZ and KVM server, which allows us to create our own VPSs and host what we want on them. So, this site is now being served up from a VPS I’ve allocated 8GB to, while my staging servers can be completely isolated if necessary.

I’ve been wanting to talk more about our tooling, infrastructure, and devops practices ( a nice grand name for a bunch of scripts), and I’m going to get prodded to do that hopefully. Anyway, welcome to the new site. I’m quite happy with it.

The post A new year, a new site and a new post appeared first on digitalquery.

Woot, I hear you all say. Woot indeed.

It’s just an interim effort, and will be replaced by a proper custom built site soon (for values of soon that range between a few months and never). I may even blog a little bit more now. Really.

Anyway, welcome to the new digitalquery site. Do get in touch if you need some social strategy consulting, some WordPress development, or a brand new community or subscription site.

The post Hoorah, a site update appeared first on digitalquery.

I posted this to the Open Melody mailing list – it turned out to be the last time I posted anything to a MT or OpenMelody group. Looking at Hacker News and seeing yet another post about a new static site generator triggered a memory of this email and it still feels relevant today.

I don’t particularly have a dog in this fight anymore – having mostly abandoned MT for new client work, but MT (and by extension Melody) is my first blogging love, so I want it to have a future.

I’ll just reiterate what my few posts to this mailing list have previously stated – Melody needs to focus on how and what to get the rest of the Perl community engaged and so that it becomes the blogging / site publishing component of choice. UI changes are nice, but, by and large, I *personally* don’t feel that this is the real battleground anymore.

Here’s what I’m seeing:

– An upswell of interest (especially amongst developers and early adopters) in static publishing engines such as jekyll and nonoc (partly driven by Github’s adoption of jekyll as the engine for Github pages.

– Despite Amazon Web Services’s recent outage, being able to serve up a website from S3 is very interesting.

– An increasing willingness to ‘outsource’ elements such as comments to the likes of disqus, intense debate and js-kit (echo).

All of the above play directly into Melody’s strengths – static publishing, template language and eliminate some of the weaknesses (comment load etc).

If I could code Perl worth a damn, I’d be making a play at the above – allowing the Melody publishing engine to just be a component, easier reading of template files directly from disk, publishing to the cloud options, direct support for popular comment services.

Having the publishing engine as a (CPAN??) component also directly brings the Perl dev community into play.

I really don’t have a sense of how easy this would be to do (my guess is not that easy or it would have already have been done), but it would result in having a coherent and competitive story for Melody, that satisfies existing needs, but also pushes Melody into a space where I feel it’s strong.

Didn’t get a single response to it, and over time, people just drifted away from OpenMelody.

Movable Type is trying to make a comeback, but apart from legacy systems and a few personal sites, I can’t imagine it’s going to be successful.

A real pity.

The post Static Site Publishing – Movable Type and Open Melody – my vision for the future from the past. appeared first on digitalquery.

Anyway, now that I’ve moved back to WordPress, and revamped the site a little, I might blog here more often. Then again…

The post It’s only been 3 years… appeared first on digitalquery.

Pinboard has been tagged as the “anti-social” bookmarking service, which is a little unfair perhaps,  the emphasis is most firmly placed on providing a simple and fast bookmarking service, with little of the increased (and probably non-rewarding) complexity of delicious.com and others.

It works (for me), it’s quick, and does what I need it to do.

The post Pinboard – a social bookmarking service appeared first on digitalquery.

First up, what he is really talking about is that RSS reader uptake within the enterprise is very low. If we narrowly define RSS reader as a separate single function application (whether desktop or browser based), then I’d probably accept that – installing new software is hard within a lot of orgs (for various reasons), and who really, apart from hard core information fetishists, needs yet another hosepipe of information ?

However, it’s a long leap to equate the lack of uptake of RSS readers (as narrowly defined above) with the death of Enterprise RSS, because the two are *not* necessarily correlated.

When I had my all too brief stint at Headshift, we (and I’m going to say we, because I think I was in the room when this was being discussed – and was definitely in the building somewhere), worked out that within an organsation, RSS was the transport layer for information, upon which all sorts of amazing social, collaborative, filtered applications could be envisaged.

And that still remains the case – from what I see, increasingly, RSS is being offered as a data transfer protocol, or a wire protocol, allowing information to be taken from one source, mashed up, munged or left alone, and reused and redisplayed somwhere else – a dashboard, a custom built application that has information display as one of its features. In many cases, no one using the system is even aware that RSS is being used , or what it is. And that’s fine – they don’t need to.

And there’s more. I’m doing some work for a PCT – a part of the NHS responsible for delivering services to local people In conversations with partner organisations, the following phrases occur often and naturally:

“We’ll supply an RSS feed of news items, you can take that and use it if you want”

“Can you give us an RSS feed of your events so we can add it to our health section”

So, RSS is being used as a transport layer between organisations, allowing simple pubish and subscribe mechanisms to move information around with little additional effort. I’d call that a very significant use of RSS in the enterprise.

So – is RSS dead ? Hell no – it just stopped being an overt feature and became part of the plumbing instead.

The post RSS isn’t dead. appeared first on digitalquery.

No recommendations yet, and not all of these are Basecamp replacments – Harvest and ClockingIt, for example, could be seen as complimentary.

• Goplan
• Unfuddle
• OnStage
• Harvest (time tracking and invoicing)
• TeamWorkLive
• ActionThis
• 5pm
• DeskAway
• Collabtrak
• ActiveCollab (self hosted)
• ProjectPier (fork of older version of ActiveCollab, selfhosted, opensource)
• ClockingIt (timetracking, opensource)
• Lighthouse
• Copper
• Wrike
• Intervals
• Central Desktop
• Huddle

Technorati Tags: basecamp, projectmanagement

The post 14 (now 18) alternatives and additions to basecamp. appeared first on digitalquery.

So the solution (where solution = not particularly pretty workaround) is to run PQ under daemontools as per this writeup by Byrne Reese. This sets up a monitor which constantly scans for the PQ job and restarts it if it’s not there.

Then create a cronjob that kills PQ every hour or so, using pkill

*/15 * * * * pkill -SIGINT run-periodic 

So, in this example, the cronjob runs pkill every 15 minutes and kills any processes that contain ‘run-periodic’  in their name (run-periodic-tasks is the perl script that runs the PublishQueue workers).

The svscan process (part of the daemontools suite) notices instantly (within milliseconds) that the PQ task isn’t running, and starts it up and memory is back to normal.

Now, this obviously isn’t ideal, but the 6A guys (and some of the more Perl aware parts members (!) of the community) are well aware of the problem, and are looking at how to fix the leaky bits, as well as make the whole thing run better and faster. In the meantime, this is working well for me.

[Update – 6A have acknowledged the memory leaks issue in the latest MT 4.1 release notes and have advised that PublishQueue not be run in daemon mode, or with FastCGI – both of which are fairly essential. My workaround is even more valid now.

Also note the addition of the SIGINT signal to the pkill command – this is a less forceful shutdown and was recommended to me in the #movabletype IRC channel]

The post Running MovableType’s PublishQueue appeared first on digitalquery.

The site was launched today – it’s a microsite for now, but will be enhanced dramatically over the coming months by the lovely people over at Splendid, and also features a blog built by me at digitalquery .

Lovely people to work with, and I’m sure this will be a really valuable resource, especially for those of us starting up.

The post Smarta launches site appeared first on digitalquery.