the rise and fall of the microsoft empire

1975-1980

Our historical timeline begins in 1975 when an unlikely duo—Paul Allen as Batman and Bill Gates as his awkward “Boy Wonder”—started Microsoft Corporation.  I’m guessing that ro-sham-bo was involved in this decision but incredibly somehow Bill was made the CEO when the company got its start.  Maybe dropping out of Harvard gives you that kind of confidence.

1981BillPaul

1981-2000

Nothing really significant happened until they managed to modify an existing operating system for the IBM PC in 1981 from another company and rename this to MS-DOS. Significant sales of the IBM series of computers and those of their competitors then launched a thirty-year stretch of dominance in the business world in the area of operating systems, software and development platforms.

For most of us, we reasonably dismissed Apple’s hardware and the MacIntosh operating systems as nothing we could seriously use in business outside of the marketing department.

Consumers bought new versions of software and that license was good for life.  It could often be transferred from one computer to the next as long as the last one was de-registered first.  If you built software for Windows, you likely used a Microsoft compiler to do so and you paid for that.  In fact, the Microsoft Technet collection of CDs was quite expensive.

2001

About six years into the “Internet Tidal Wave” as Bill would call it, Microsoft was starting to lose its way.  They tried to dominate in the browser wars but never quite managed to quash the competition.  Others saw their efforts in this area as annoying.  Their software for creating programs, Visual Studio, first hit the scene about four years prior to this.

Google was founded some five years prior and was just beginning to get attention from an investor before they had anything real yet.  In 1999 they moved from their garage to an actual building in Palo Alto.  Yahoo’s popularity as a search engine from a decade ago was waning.  Google’s ad-based revenue from keywords was paying off; they’d planted a money tree which eventually created an entire forest of money trees for them.  It wouldn’t be long until Microsoft’s executives behind closed doors would consider Google their biggest threat.

About this time Apple created a very clever method of provisioning content for one-and-only-one device within the music-delivery space.  The iTunes store would turn out to be the goose that laid the golden egg, as seen in the following revenues.  And yet, it would take years for either Microsoft or Google to realize the beauty in this fulfillment model and to come up with their own versions.

showmethemoney

The “Internet of Things” concept started gaining in popularity at this time.

2009

Microsoft’s attempts at copying Google’s success (MSN Search, Windows Live Search, Live Search) now culminated in the introduction of Bing as their default search engine destination for all things Microsoft.

Apple introduced the first iPhone and the first iPad about this time, noting that the same provisioning model from iTunes was incorporated into both via iOS.  The subscription model of sofware licensing was born with this, if you think about it.  If you wanted to write a program for either, you needed to use Apple’s software to do so.

Google has just introduced Chrome as a browser and would begin their campaign to slowly break Internet Explorer.  The same was true of the Android phone and its related operating system.  It would take a few years for Microsoft to catch up to either the iPhone or the Android before releasing their own app-savvy smartphone offering.

Amazon some three years prior had introduced the beginning of what would be a full complement of cloud-based services to support web development.  It would take Microsoft two full years to realize that they needed to be in this space and they didn’t have their offering ready for a few years more, too late to effectively compete.

Github.com had just celebrated their first year online, hosting over 46,000 repositories by then.  The world of open source was the very antonym to the way that software had been developed prior to this.

The free Ubuntu operating system was released about four years prior to this, backed by the well-funded company Canonical.

2015

Microsoft releases Windows 10, “the last version of Windows” (they claimed).  Rumors suggested that Windows would eventually go from a version-based license model to an annual-subscription model with respect to pricing.  I think it’s safe to say that the market hasn’t really embraced either Windows 8 or Windows 10.

The subscription-based model for Office 365 was introduced four years prior to this so the writing was definitely on the wall:  Microsoft wanted to depart from their former methods of making money and to chase the monthly subscription model.

2015-popular-coding-languages

The world of open source was offering new programmers a wealth of free code.  All they had to do was to take it and make it their own.  Formerly, Microsoft-friendly coding languages like C, C++, C#, VB and .NET dominated the playing field but this graphic shows how the game had changed.

2017

And here we are, present-day.  That curious number 42 now describes the number of years that Microsoft has been around.

Yesterday evening, I attended a very geeky meetup of perhaps fifty or sixty coders and only saw one Windows-based laptop.  Almost everyone had a MacBook of some kind.

I just spent about two hours today installing the free Visual Studio Community 2017 software so that I could—in theory, anyway—alter a free copy of the source code for TightVNC software.  Out of the box, so-to-speak, Visual Studio doesn’t want me to build this project since it uses an earlier target platform (Windows 7 or 8, one would assume).

Microsoft only wants me to make things for Windows 10.

So rather than making it easy for me to build a program that will happily work with Windows 7, they’re forcing me to jump through hoops in order to add the necessary pieces for this to happen.

Add two more hours to this and I find that my installation does not want to download the earlier pieces to allow this to happen.  I’m forced to then upgrade the code to Windows 10 compatibility mode… only to find that the build fails with 528 errors.

The main crux of all these errors appear to be:  “we can’t find common files”.  It’s a very amateur sort of error from a company that’s been providing compilers for several decades now.

I have to think that Microsoft doesn’t want me to do anything with Visual Studio unless it benefits Microsoft.  And this is the core of the reason why I suggest that they’re doomed.

Every time a coder like myself runs into obstacles like these, the usual seed that’s planted inside their head is “this would be easier with another free compiler or another language from someone else”.

2022

Fast-forward another five years and Microsoft will have lost ground on many fronts.  New software development here, there and everywhere will be via some language which wasn’t popularized by Microsoft on computers which aren’t Windows and with browsers which aren’t Internet Explorer or Edge.  Our toasters and refrigerators and our cars will be powered by the Ubuntu operating system or perhaps Debian, a similar free Linux flavor.  These appliances will be connected to our wi-fi and even to the Internet but there won’t be a scrap of anything Microsoft about them.  They’ll be coded up with something that isn’t C#, doesn’t use .NET and doesn’t need Visual Studio in order to compile it.

The only thing with a Microsoft pedigree with some staying power could be some of the websites and services currently served up at Microsoft’s datacenters via Azure.  But Amazon or Google could kill that by simply lowering their own prices for cloud-based services.

iphone without itunes

You know how Apple can be sometimes; they feel the need to control everything. So for a Windows-based computer, they want to force you to install the entire iTunes collection of software just so that you can get to your files on your iPhone. As an I.T. person, to me that’s just way too much software to be adding to someone’s computer setup.

Why not?

You might just ask “why not?”  Why not just install iTunes? One of the subtle changes that iTunes makes in terraforming your Microsoft computer for its own needs is to install a variety of software to make things more Apple-friendly.

For example, in an Apple-based network the Bonjour service allows lookups for printers normally but allows for almost any device to broadcast its existence on your network. The downside to adding a different printer lookup service is that you might have a number of printers already which broadcast via Bonjour and can now be seen by your computer this way.  And yet, you might not have a working Microsoft driver installed to make all this happy. The printer when added simply doesn’t work and yet it seems to work for everyone else on the network who didn’t install iTunes. Rule of thumb for success: don’t arbitrarily add services and things unless you exactly know the ramifications for doing so.

Rule of thumb for success: don’t arbitrarily add services and things unless you exactly know the ramifications for doing so.

The problem

If you simply plug in your iPhone into a Windows 7—based workstation you’ll see it download and install a default driver. Unfortunately, the Internal Storage section of this device won’t show anything in it.

iphone-no-driver-yet

The fix

Unbelievably, the fix is much easier than I’d imagined. Immediately upon tethering the iPhone the very first time to the Windows computer the iPhone will buzz twice (telling you not that it’s now charging but it’s trying to tell you that it’s displaying a notification).  The message is crucial to your success but Apple in its infinite wisdom doesn’t decide to wake the phone up for you.  You need to manually wake it up first to see it:

allowthisdevice

Select the Allow option here and suddenly Explorer will now present you with a DCIM folder, below this a 100APPLE folder which contains your images.

iphone-after-allow

Why is this considered a smartphone?

That’s a good question to ask. Why would Apple decide to block access to the phone on a Windows computer by burying its head in the sand when an important access message is being hidden behind a sleep state? I suppose they could suggest that if the phone is sleeping then the rightful owner may not be in control of it and that nobody should have access as a result.

But why not simply bubble that information up to Explorer with a dialog box so that the user will know the status? It just silently doesn’t see anything at all for the device.

If you read the many support threads on the Apple site nobody ever mentions such an easy solution. The reason of course is that Apple wants you to install all of their software on your Windows-based computer, too. The biggest reason is that the iTunes application is a shopping cart and you’re a consumer to them.

ubuntu core on the raspberry pi 3

The fun just never stops here as I begin with a new operating system on a second microSD for my newly-purchased Raspberry Pi 3. This time, it’s Ubuntu Core. This operating system from Ubuntu appears to be a big departure from the Ubuntu server or desktop versions I’m used to. In case you were wondering, I managed to install the new operating system with little trouble. It’s different, though, I’ll be the first to admit it.

Goodbye apt-get, hello snap…

Gone is the usual apt or apt-get interface for fetching code. It’s been replaced completely by snap. Honestly, apt has been the mainstay command for managing Ubuntu (and Linux) for a while now so this is quite the departure from the norm. Snap, in theory, will be a cross-Linux way of deploying code. From my initial research, it appears to be a lot like the Juju Charms for deploying services on a Metal-as-a-Service (MaaS) system. Like Charms, Snaps appear to have an up-stream and down-stream component to them, providers and consumers, if you will.

On the positive side of things, a Snap that you create would work on many different systems. In this way, it works a bit like Java or .net (somewhere, there’s an interpretive layer for the particular hardware but you don’t have to worry about that). Presumably, you mostly just worry about the interface you’re providing and the packaging requirements to create a Snap for submission to the store.

Since Snaps get digitally signed this makes them more like the Apple iTunes store metaphor that Microsoft, Google and everyone else seems to be going with these days. Digitally signing your programs both lowers the risk of rogue/evil code out there but it also puts a middleman into your money stream—someone like Apple will be there to charge the publisher a fee for making money on the store and for signing your developer’s certificate each year. At the moment, Canonical (the maker of Ubuntu) doesn’t charge for a Single-Sign On account but maybe if this becomes popular they will in the future.

Honestly, the entire concept of Snaps appears to be a watered-down clone of Apple’s iTunes distribution model.

Fee Structure for their Store

Not that this information is easily available, I managed to finally find it. From this page on Canonical’s website:

4. Pricing, fees and payment

  1. If you elect to distribute an App without charge, the payment terms of this Agreement will not apply with respect to the free App.

  2. If you set a price for your App we will collect fees from end users that purchase your App at the price you set. Within 30 days of the end of each calendar quarter, we will provide you with a report of the number of copies of each of your Apps sold and the amount of any payment due, which shall be the fee multiplied by the number of copies sold less any applicable taxes and our commission. Our commission is 20% of the total fees charged for the sale of your App, less any applicable taxes.

So they’re charging 20% commission for brokering your Snap. Compare this to Google’s flat 30% commission, Apple’s flat 30% commission plus $100/year developer license and Microsoft’s 30% commission plus $49/year developer license. Note that there are precious few Snaps available. Specifically, only one Snap has a price of $1 at this time and the remaining 547 are free. So until others are charging for their Snaps, don’t expect to make a dime selling yours, in other words. In a free world, nobody will open their wallet nor expect to.

The Future of Snaps

It’s difficult to say whether this entire concept will get traction. Linux has been the free alternative. The people who run Ubuntu, for example, like that aspect about the community. Will these people easily change their stripes and embrace a payment system in the future in which we pay for code? Remember, we’re talking about people who’d rather build from source instead of paying for a binary file. My gut tells me that they’re never going to go for it.

Should I Invest Time in Developing Snaps?

As a developer, having a variety of experience looks good on your résumé so yes, by all means, develop a Snap. And since we developers often communicate our value via our github page I’d suggest that these Snaps be free and open-source if possible.

a new app pricing model

Personally, I don’t like the idea of “in app purchases” which seems to be the norm these days in the current app pricing model. You get the app for free, try to use it and then find out that you can’t save your results unless you pay the author for this feature. If you’re like me then you usually feel like they’ve wasted your time.

I would suggest that a better model is how we now buy music.

Changes in the recording industry

You used to have to purchase the entire album just to get a single song, that was just how things were. If an artist or band had one killer song that was enjoying a lot of airplay on the radio and assuming that you really wanted that song then you had to pay the $12 or so for the entire album. And you just crossed your fingers that one or two of the other songs made it worthwhile.

With the advent of iTunes and similar websites, we now have the ability to sample and purchase exactly which songs we want to pay for. If half of the album isn’t worth it, you don’t have to buy all of it. If the artist only has one good song then there’s no reward for them to pad the album with a lot of junk.

A new model for app pricing

So why shouldn’t we just show our app’s prices up front instead of hiding them inside? Currently, a new customer can’t see how much of the app is crippled and how much is functional until it’s been downloaded and used. How many times have you downloaded and demo’d two or even three different free apps of the same kind, trying to find one that was reasonably useful?

What we need is a venue for selling our apps like musicians sell their songs. Theoretically, it might look something like this:

PayAsYouGo_734x466.png

More feature transparency

The biggest benefit to a model like this is that it shows the potential customer what’s included in the full program and the cost of each feature. If, like me, they’re not interested in the social connector feature then they simply don’t purchase it. You pay for what you need and nothing more.

As developers, we would distribute modules of functionality and charge the user on a per-module basis. I would suggest that features be priced differently based upon the perceived value. In fact, there’s nothing to prevent the price of a popular feature from increasing over time.

Like in the iTunes model, clicking a play button next to the feature ought to bring up a demo or screenshot of the feature in action.