Thoughts and Rants on Software Design

I loathe having to call up any big company to get service.

Some parts of the process are just reality but many could be massively improved with a few small changes.

In no particular order here’s my beefs:

1.) Volume control – if your “your call is important to us” voice recording is so quiet that normal people can’t hear it… or so loud I can’t hold the phone to my ear – you’re annoying customers before they can even talk to you.

2.) If the Customer Service Rep (CSR) will be unable to do anything without my account number, name of first born, and the last 6 digits of my credit card – make sure you inform me up front and collect that data from me while I’m waiting for a representative.

3.) Those automated systems that ask callers to verbally state which option they want from a menu are officially the most annoying automated systems ever designed! Seriously if you can avoid ever installing one of these your customers will thank you. I don’t think they have ever correctly grasped my intentions. “How can I help you?“… “Tech Support“… “Did you say [Printer Ink Refills]?“… “No!“… “Let me try again – how can I help you?“… “I want to talk to a real person!“… “Did you say [Wifi Setup]?“…”WTF!? you stupid piece of ****!“… “Did you say [Install Anti-virus]?

4.) Most times when you are calling for some level of tech support the CSR has a script they want to run through to resolve your issue.  It would be soooooo much better if they would at the beginning of the call politely ask what your level of technical expertise you have (or listen when you try to tell them!) so that you could skip over some of the endless steps and explanations (e.g. I know how to bounce my router, get & release my IP address, my version of Windows, how to open Device Manager, etc.)

Behind the scenes of course I realize that not every company has a good system to take care of all the details… and that the staff that know the products/services very well are likely not working the phones but anything a company can do to improve this workflow would be really appreciated by customers and hopefully happy customers will improve your companies bottom line. :-)

We’ve had some pretty rough weather recently – an ice storm, snow, and a cold snap that made the North Pole feel like a warm place to visit.
As a result trees snapped and fell causing power outages (some for days!), roads were blocked and some odd things occurred.
Anyway on my travels this weekend took me to an industrial area where I heard an alarm sounding. I didn’t have to investigate but I did.
There wasn’t any signage on the building so I drove around for a closer look… and here’s what I found.

I filmed the above expecting to send this to the local fire department or similar to investigate. I went back around the front of the unit to jot
down the address details and I ran into a guy that worked there. It turns out that the sprinkler pipes broke and the flooding inside was as bad
as expected… he called the sprinkler guy but there was likely another 20 minutes of flooding before they got there :-(.

I’ve still got no idea who the company is, what product they make but I can only imagine that no matter what it is the water damage was extensive and expensive.

What’s the moral of this story?

Well it all ties into programming and code. I saw an issue that was otherwise being ignored (the alarm was loud and could be heard from quite a distance by many people).  Rather than just ignoring it like others were I made an effort to investigate and resolve the issue – I couldn’t as a good citizen just let the alarm go unanswered.

The same goes for code and being a good citizen programmer. If I find an issue in code (bad logic, memory leak, just plain wrong code) I don’t have to be the programmer that fixes the issue (I doubt I would have been able to put out the fire (if there had been one) nor stop the flooding) but I did need to do something!
On the code level I could also just file a bug report or told someone else about it so that it could get the attention it needs.

If you are looking to be a better programmer, be a better citizen (or a combo of both) – the next time you see something ugly in code you’re looking at – don’t ignore it… either fix it or report it.

Anyone that has talked to me knows that I’m not a huge fan of Windows 8 by any stretch of the imagination.  However its a bit unfair to turn my nose up at the whole operating system (especially when there are some really great parts!) so I thought I’d break down a bit about what I don’t like.

Today’s focus is on the scrollbars in Windows 8 (and back-ported to Windows 7 inside IE10/IE11).

Scrollbars are one of the necessities of an OS to let you get to content beyond the visible container you’re looking at… and each part of it serves a specific purpose.

The arrows at each end are “buttons”, the part that slides up and down (or left to right) is the “thumb”, and it slides along the “track”.  There’s actually been a bunch of advancements with scrollbars since they were first used.

For starters the size of the thumb used to be fixed (same as the arrow buttons) but by making it proportional the user gets a visual clue as to how much content there is to scroll and subsequently makes scrolling more controllable and easier (bigger target, easier to hit a la Fitts’s Law).  Secondly when horizontal and vertical scrollbars are used together there is a little box in the (typically) bottom right corner where they meet that is empty.  Some smart developer felt this was a good place to put some diagonal lines (often referred to as a grippy) to enable a user to grab the corner and stretch the container/window to a new size.  Finally although one can debate the true usefulness of it, clicking in an empty spot on the track will scroll the container to that position.  This helps a lot when you have a 500 page document and you want to quickly scroll to 300-400 pages down.

So here’s a few screenshots of a typical scrollbar in Windows 7:

Windows 7 default scrollbar

Windows 7 default scrollbar

and here it is being moused over


Windows 7 scrollbar moused over

This interface feature has been refined over the years ever since Microsoft introduced Windows 3.x such that now the bar is a nice soft gray color with slightly rounded corners and with shading, highlights and a gradient that gives it that 3D look that literally makes you want to bring your mouse over and interact with it.

The control supports visual states for “normal”, “hover”, & “click/drag” and was IMHO pretty much perfect, better than Windows XP, 2000, 98, 95, etc.


Now of course Windows 8 comes along with their “flat” UI to try and bridge the “simple” tablet UI with the desktop.  This is no easy task and well I fear that with a “compromised” tablet and desktop interface neither extreme is really ideal and everything became a muddled sacrificed UI.

Here’s the scrollbars in Windows 8:


Windows 8 scrollbar missing hover / mouseover state

In this shot you can see there is no hover state for the arrows and all the graphical details, soft ergonomic curves that afforded discover-ability are gone. e.g. “sexy-factor of zero”



Windows 8 scrollbar arrow is lopsided

When actually clicked the arrow looks peculiar like it wasn’t built properly… the arrow is not centered as one would expect and the result looks like a graphical glitch.

I think it would look much better if the arrow was centered on the “button” like this:


What the Windows 8 scrollbar arrow should have looked like

I’d still argue that these scrollbars (especially on the desktop) just seem completely out of place and look ugly.  It reminds me of using Balsamic (a tool I love) or similar to quickly mockup UI screens where there is an intentional effort to make it look rough and unfinished… only Microsoft actually shipped this!

What’s your thoughts? Do you like the Windows 8 scrollbars? Do you think they should support a mouseover / hover event like all the other controls? I’d love to hear your thoughts.

Full disclosure – this post is also a bit of shameless self promotion.

I’ve had an idea for a website / webapp for quite some time and I’ve recently been working on bringing it to fruition.  I’ll be chronicling the progress here on this blog.

Due to the nature of the Internet, Startups, and the importance of being “first” to a particular domain I’m going to be keeping a bunch of details under wraps until ready. ;-) However there’s a few details I can share so here goes.

I’ve always been a big music fan and love to share my passion with other music buffs whenever possible but I’ll concede that outside of a one on one chat with a friend I haven’t found a good online experience to share this energy.

Well with this new site I hope to change that.  What if there was a site where you could share the knowledge and experiences you’ve had with music? Do you remember singing along to “The Power Of Love” in Back To The Future… or what band sung the theme song to Friends… or what song is playing in the background of the latest Doritos commercial?

Maybe you are the local expert on all things U2, Britney Spears, Garth Brooks, Psy, One Direction or Daft Punk? If you’ve into music at all you’ll likely enjoy what’s coming with SongKiwi and should sign up for more information and to get on the beta invite list. I promise it will be both fun and useful.

Signup for information on the SongKiwi beta.

I’ll try and give a bit more info in posts to come especially when a bunch of the technical details have been decided upon. HTA6UNSTYW9W

This isn’t going to be (or at least it certainly isn’t intended to be) a fight-inducing post about native-vs-HTML5 apps on mobile… but rather something I’ve noted as a bit of a flaw of the new Native Mobile App model where you download an app from a specific App Store/World/Marketplace.

For those of you that have been around the Interwebz since the beginning (or at least say… circa 2000) you’ve likely noticed that most applications you download for your computer (PC, Mac, or Linux) have gotten faster and faster as the years go by.

There’s 3 major reasons for this:

  1. Internet connection speeds have massively increased from our measly 56k dial up modems
  2. Use of CDN (Content Delivery Networks) has improved things so that you download from a server that is geographically closer to your location
  3. Vendors now provide a quick download installer that fetches the required resources only when you start installing the application

Item #3 is the key.  When you download the latest version of say “Mozilla Firefox” you get a tiny installer that downloads in seconds.  Only once you start it up does it do a quick inspection of your hardware (to find out if you need the “Windows 7 32bit” files, or the “Windows 8 64bit” files for example) and maybe asks if you want language packs, extra features, etc. then downloads only the assets you need for your installation.

The same applies whenever you want to get an update! You download the new parts that you need and essentially “patch” your existing installation.

This is where the flaw comes into play with Mobile Apps downloaded on your smartphone or tablet.  The publisher of say “Angry Birds” (Rovio Entertainment Ltd. btw) (weighs in at: 39.9MB) might have fixed a typo or a silly off-by-one error amounting to less than 100 bytes… yet every user that wants to get that update has to re-download just shy of 40 MB of data!

Not only is this a major waste of network bandwidth but if you are applying these updates over your 4G cellular network while whipping down the highway at 75 miles/hour (120km/h for the metric folks) you’re not only putting unnecessary stress on the cellular network, and your battery… but brutally eating into your bandwidth cap (for those that have them) or (gasp! roaming data charges!). You should be saving that precious bandwidth for watching funny cat videos! :-P

This feels like an epic failure to me to have come so far with online software delivery… and then just throw it down the toilet and re-download entire apps whenever there is the slightest change.

Now I do realize this is quite a challenge to solve as you’d need to provide a smart solution that accounts for literally 1,000′s of developers coding styles/structures.  Some sort of magical binary diff… but that’s what motivates us developers! A Challenge! Announce: “It can’t be done!” and dozens of developers will stay up all night to try and prove you wrong!

Ok, time to play Devil’s Advocate

  • I guess if you had a hybrid app with a small native wrapper that loaded most of the content from a remote server on install/access and cached it as “app related files” then you can overcome the “big download hit” for updates.
  • HTML5 advocates will point out that this is how & why the HTML5 appcache was designed.  Enable smart caching of the resources that can be cached and used offline and only fetch fresh resources as/when needed.
  • A native app with a collection of simple WebViews could combine the benefits of the above 2 items for an app that might never need updating yet always be up-to-date.
  • I’m sure there are a bunch of bright developers that have already taken one or more of the approaches above (even for completely different reasons) but serendipitously solved this problem.

What’s your thoughts? Are we messing things up? Will it all not matter the closer we get to wireless “Gig-E”? Or will the big mobile vendors (Apple, BlackBerry, Microsoft, Google, Mozilla) come up with an ingenious solution?

Scary Code – Red Flags

Every time I navigate through code there are things I see that I would call “Red Flags” that are indicators that either the previous developer(s) didn’t know what they were doing or they built a fragile system.

Here’s a few of the one’s I’ve seen that scare me (and an explanation why if it isn’t obvious). Sadly there are many so I suspect this will be a recurring thread ;-)

Red Flag Warning

1.) Anytime I see a variable or parameter named something like “ignoreErrors”.  This could be a totally legit case… but more likely points to either depending on code that often throws random errors or simply not dealing with errors properly and just trying to provide a duct tape solution.

2.) Anytime I see “temp” in a variable name. Again there might be a rare case when shuffling 2 indexes in an array but any other scenario suggests that the developer didn’t spend the time to properly name things and may well have not spent the time to write good logic either.

3.) Anytime I see a catch statement with nothing in it.  If commented with an explanation of how failure is ok then this is fine.  However any other time I see this it looks like the developer left the stub in there to handle later but never came back.

4.) Anytime I see documentation for a method that exceeds the size of the method itself.  This is typically a sign of code master… uhm… what’s that word for when you have the ultimate fun by yourself?  This is a sign of a developer that is actually thinks so much of themselves they feel the need to advertize their skills by going on and on about their code.  More importantly code should be self documenting so if you need to add huge blocks of comments to document your code something has gone wrong.

5.) Anytime I see a method name that ends in a number “2″ (or worse yet a 3, 4, 5…) e.g “getBillingInfo2″  This is a sure sign of an initial method not quite being up-to-snuff but rather than the developer fix the method to make it more robust (or provide overloaded/helper methods) a previous developer was lazy and simply cloned the original and hacked in their own stuff.  Other classic symptoms of this disease are methods prefixed/suffixed with “Old”, “New”, “VersionTwo”, “VersionThree”.

Got any “Red Flag” warnings of your own? Let me know and we can share the painful burden of maintaining bad code!

Programming Rule #1

The counterpart rule to programming rule #7 is this one.

Programming Rule #1

“Thou shalt only load the data you need, and only when you need it.”

This is a rule that affects everyone as an end user.  Users can see exactly what content loaded on the screen and when it takes longer than expected the experience is ruined.  As a developer every time I see inefficient code that is loading more than it needs to… or well before it is potentially needed or disastrously even when it is not needed – I cringe.  Its a waste of resources & network bandwidth and it makes screens load slower affecting not just the end user but every developer or tester than needs to wait for that page to load.

Of course the exception to this is when you have a ton of data that makes more sense to pre-load.  That’s fine too (and covered in programming rule #7) ;-)


Get every new post delivered to your Inbox.