Category: Software
9 Things Cullect Taught Me About Software
- Forcing people to create an account to use your software is a bug.
- if you’re not scared to deploy, you’ve stopped caring.
- Murphy is alive and well.
- Google and a bookself of technical books can be equally useless.
- Good software is like an iceberg.
- if you ask for money, people will give it to you.
- Software is as hard as you make it. Don’t.
- Most features can be removed and no one will notice.
- A great project will eat everything in it’s path.
Planting Flying Meat Acorn Near Photoshop Elements Grave
Like most professional graphic designers, my career was measured in versions of Adobe Photoshop.
v2.5: I decided I wanted to be a graphic designer. The dad of a high school classmate was one. I went to talk with him about it. He worked out of his basement home office with a view of the lake, a room full of Apple gear, and telecommuted to Minneapolis. He launched Photoshop and showed me some of the crazy stuff he was doing with it. I left stunned.
v3: Spent far too long on the Photoshop Classroom-in-a-book tutorial as part of my least favorite college course – something about printing methods and preparing images for press.
v4: The first version I used in a professional environment auch auf Deutsche.
v5: Editable type, Multiple Undo. Those two features are reason enough to fall in love with it.
v6: Like Word 6, and Star Wars 2 – all the hope, promise, performance, and love of the previous version was an unsaved memory. This is the last full version I used and the beginning of my strained relationship with Adobe.
In 2002, my new digital camera shipped with Photoshop Elements v2 and then a couple years later, my scanner shipped with Elements v3. Aside from a few small omissions (inverse selection, select color range, etc), Elements was all the image editing horsepower I needed. It was my go to app, until I switched to the MacBook.
Elements never launched on the MacBook. It’d just bounce and bounce and bounce and bounce until I forgot why I opened it and Force Quit (a pretty good indicator of my image editing workload over the past 2 years).
After 15 years, my relationship with Photoshop officially ended today.
A while back, I downloaded Flying Meat’s Acorn and hadn’t opened it until this afternoon. While Elements was bouncing, I opened Acorn to take a closer look at a client website mockup. Instinct kicked in and I was pushing pixels, using the same key commands I remembered from Photoshop.
Before I hit Save the first time, I bought a license.
In addition to costing less than 1/10th the price of Photoshop, it was the most integrated web/desktop licensing experience I’ve seen. After completing the purchase online, a single click in the browser applied the license to the still running desktop app. Seamless. Fast. Amazing.
I’m one step closer to being Adobe-free and happier than ever.
Google Gears and Joyent Slingshot – Take the Web Offline
My biggest issue with browser-based apps is their complete uselessness when I’m without internet access (or low bandwidth situations). Still happens quite frequently.
I knew that situation wouldn’t last long, but of all the potential outcomes, I didn’t expect this one: Joyent Slingshot and Google Gears turn browser-based apps into offline desktop apps.
Crazy.
When Not To Do a Holiday Logo for Your Software
Earlier this week, graphic designers everywhere swapped out regular logos for Halloween-themed ones. Google, MacUpdate are just two I bumped into within my browser.
Outside of my browser – TextMate – also changed it’s normally non-descript logo earlier this week to a glowing jack-o-lantern.
The difference is huge.
Each day, I ignore Google’s logo microseconds at a time. It’s out of the way and I’ve been trained to use their page layout and CSS to identify ‘Google’. Same, but to a much lesser degree, goes for MacUpate. Web services can mummify their logos, because they’re like name tags at a conference. Nice to have, but after a while – completely useless.
Changing the logo on my paid-for, always-on, desktop software impacts my productivity. It actually slows me down by requiring me to think longer about what I’m doing rather than just do it.
Questions I’ve asked since TextMate changed their logo:
- Is TextMate open?
- Where is TextMate?
- What’s this pumpkin application?
- Where is TextMate?
- When will the icon revert?
- Why hasn’t the icon reverted yet?
- Man, this is annoying.
- What was I doing?
All of these questions take attention from what I’m doing, and put it on TextMate. I’m on the Mac to eliminate applications begging for my attention. Speaking of Apple, if you’ll recall, iTunes has tweaked their icons nearly with each new version – the extent of this change: a different color musical note.
Update 2 Nov 2006: [REVISION 1324] made it all better. Thanks TextMate.
Free and Open vs Not – At a Glance
Steve Borsch weighs free & open against for-fee and closed. In in he brings up some great points – namely, if the problem you have is solved – good enough – by a hosted, for-pay service, then installing and setting up a “free”, open-source system isn’t worth it.
Personally, I’m not keen on SurveyMonkey’s presentation – and if I had need for a browser-based survey, I’d want to polish the presentation more than they permit easily. In that sense, tweaking PHPSurvey might be worth the effort. Same may be true of integrating into other systems. Same may be true if I, for whatever reason, don’t want the service provider to have my data.
Whichever solution I go with, it will take some amount of setup time, time to get familiar with the tool, and time to make it work the way I work. Question is – which will make the most sense for my specific problem.
Depends. If the problem doesn’t include customization or integration, then open source isn’t a good candidate. If the inverse is true, then an open source project will get you up and running faster than building from scratch.
I do agree with Steve, there’s a huge opportunity for organizations to take free, open-source projects, polish them up, make them dead-simple for a specific group of people to use, and sell access to the implementation back to the audience. Stikipad v. Instiki as wiki solutions come to mind. This transformation:
- Is dependent on the open source project’s license (GPL doesn’t allow this, MIT license does) and
- No longer makes it a free and open source project – it’s then a commercial product. Hopefully contributing back to the original project.
Applications and Features I’m Interested In
- Shared read/write calendar – so multiple people can change the information on a specific event.
- Shared presentation application – so a group of people could easily see the same thing, controlled by a single person, sans-projector.
Advice to Web Developers: Forget the Password
This weekend while wandering down the aisles of our local Super Target, we found a dinner table and a side board we though would go great in our living/dining room. After checking out, a couple of teenage boys wheeled the still flat-packed pieces to our awaiting PT Cruiser.
Now, after flattening the inside of the car, both pieces fit. Though either Jen or myself wouldn’t. We kindly asked the boys if they could hold the pieces until I returned.
Sure.
After dropping Jen and the little man home, I returned to pick up the furniture – now in the Customer Service area.
“I’m here for those pieces.”
“Do you have the receipt?”
“No.”
We chatted for a bit, trying confirm that the pieces were in-fact mine and paid for sans receipt.
I told her we couldn’t take them before, because we couldn’t get them both in the car.
She called over the same teenage boy and off we went.
One of my bigger irritations these days is with the number of passwords I need to remember to try out the latest browser-based Web2dotOhGodNo beta.
Frequently, there’s no real need for a specific web service to require registration of a unique identity, let alone I’ve already generated a pile of them elsewhere (can’t I use one of those?).
Sometimes, my browser will pre-populate the login/pass – that’s great while at the same time completely defeating the purpose of security. Security and identity are separate concepts, though security may confirm identity, there are other ways.
Point is the two concepts are mixed up so much there’s an inherent security problem.
The more passwords I create, use, manage, and remember on a regular basis, the greater the chance I’ll use something like “1234” and the whole ecosystem becomes insecure.
I’m using Apple’s Keychain Access to store passwords both me and my browser have since forgotten. Passwords for trials that have expired and services that no longer exist. Thing is, I’m far less likely to click ‘forgot password’ than I am to never return (Who knew Friendster was still around?).
Forget the password, it’s a security risk for customers and a barriers-to-entry for providers.
ELSEWHERE:
Software Distribution History: Shrinkwrap to Download to Appcasting
NetNewsWire, my preferred RSS reader, isn’t particular about the file type within a given podcast. Audio (podcasting), video (videoblogging), images, pdfs (like 101sheets), torrents, or even applications (appcasting?).
As you can tell from the appcasting link, Fraser Speirs was the first I knew of using an RSS feed to distribute his excellent iPhoto Flickr plugin. More recently, the new version of Coding Monkeys’ SubEthaEdit came through their News feed.
Brilliant, this makes RSS 2.0 is the universal format for distributing updates of anything. Ultimately, I’d like to see the SubEthaEdit feed integrated into SubEthaEdit, same for all my other apps. Then we can get rid of those ever awkwardly implemented ‘Check for New Version’ menu items.
Bonus link, WP-GotLucky, another WordPress plugin I spun together, turns Google referral queries into an RSS feed. Making search engine performance more real-time and more visible than my server log analysis program supports.
A Better Dircaster for the FastCast
In my quest to make podcasting as easy as possible, I’ve hacked the dircast (turns a directory of mp3s into a podcast) to support getID3() (reads the ID3 tags; artist, album, comments, from an mp3 file). I’ve also cleaned up the code a little, making it easier to customize.
Download BetDirCaster from here
The first project using BetDirCaster is my fastcast experiment. In contrast to the long, conversational form of First Crack Podcast, fastcast will be more like voicemail. Very quick, unedited thoughts, under a minute. Messages rather than conversations.
I’ve set up Audio Hijack Pro to write the ID3 tags, when I stop the recording, Automator tells Transmit to upload the mp3.