Files exist on PCs but not in the cloud

We're quite familiar with the concept of data files and file systems. And why shouldn't we be? We've been using files to store data since the first PCs and PC-DOS, and even earlier than that. (CP/M and other predecessors to PC-DOS used files and file systems.)

But computer systems did not always store data in files, or provide a familiar API into the file system.

Prior to the PC revolution, there were non-PC systems used for various purposes. Mainframes, minicomputers, and a rather interesting device: the dedicated word processor. These word processing systems were small (ish) computers that were built for one purpose: the composition and editing (and eventually printing) of documents. They were larger than today's typical tower desktop PC system, and even larger than the original IBM PC -- but not much larger, and they fit comfortably in many offices.

While they were computers, they were single-purpose computers. One did not choose to run an application from a list; the only application was the word processing software. (Quite similar to today's Chromebooks which run only the Chrome browser.) They were made by different vendors (DEC, Wang, and even IBM) and each was its own separate island of computing power.

These dedicated word processors did not have a "Start" menu, they did not have Windows Explorer, they did not have a command-line interface. All of the "common file operations" that we associate with Windows Explorer were handled inside the word processing system, usually in a "utility" menu. One did not work with files but with documents; one created a new document, edited a document, or printed a document.

Documents had names -- nice, long names which allowed for spaces. Since this was a computer, and since it did have a file system, the word processing system mapped the document names to arbitrarily assigned file names. The user never saw these file names; the word processing system insulated the user from such minutiae.

These dedicated word processors stored your information and retrieved it for you, and you had no opportunity to work with the data outside of the system. Unlike personal computers (with data stored in the file system accessible to other programs), word processing systems shielded the details from the user. Users never had to worry about file formats, or parsing a file into another format. The word processor was the only program to access the data, and the user ran only the word processor. Your data was a captive of the system.

Those dedicated word processing systems with their captive data are things of the past. PC-DOS, MS-DOS, and Windows all use file systems -- even Linux uses file systems -- and these file systems allow us to always access our data with any program we choose. We can always examine the files which contain our data (provided we know the location of the files). What does this ancient history of computing have to do with today's technology?

I think that this shielding technique could make a comeback. It could be revived in the world of cloud computing, specifically in the Software-as-a-Service (SaaS) offerings.

Anyone who has used Google's "Drive" feature (or Google Apps) knows that they can create and edit documents and spreadsheets "in the cloud". Using a browser, one can create an account and easily work with one or more documents. (Or spreadsheets. Let's call all of these items "documents", instead of the verbose "documents, spreadsheets, presentations, and other things".)

While one can create and edit these documents, one cannot "see" them as one can "see" files on your local hard drive. One can manipulate them only within the Google applications. The PC files, in contrast, can be manipulated by Windows Explorer (or the Mac Finder, or the Linux Filer programs). A document created in Microsoft Word and stored on my local drive can be opened with Microsoft Word or with LibreOffice or with various other programs. That's not possible with the documents in the Google cloud. (I'm picking on Google here, but this concept holds for any of the Software-as-a-Service offerings. Microsoft's Office 365 works the same way.)

Documents stored in the cloud are documents, but not files as we know them. We cannot view or edit the file, except with the cloud provider's app. We cannot rename them, except with the cloud provider's app. We cannot e-mail them to friends, except with the cloud provider's app. We cannot see the bytes used to store the document, nor do we know the format of that stored form of the document.

In fact, we do not know that the document is stored in a file (as we know the term "file") at all!

The Software-as-a-Service systems do allow for documents to be uploaded from our PC and downloaded to our PC. On our PC, those documents are stored in files, and we can treat them like any other file. We can rename the file. We can open the file in alternate applications. We can write our own applications to "pick apart" the file. (I don't recommend that you devote much time to that endeavor.)

But those operations occur on the PC, not in the cloud. Those operations are possible because the cloud system allowed us to export our document to our PC. That export operation is not guaranteed. I could create a cloud-based word processing service that did not allow you to export documents, one that kept your documents in the cloud and did not permit you to move them to another location. (Such a service may be unwelcome today, but it is a possible offering.)

The ability to move documents from cloud-based systems is possible only when the cloud-based system permits you to export documents.

Even when SaaS systems allow you to export documents today, there is no guarantee that such capabilities will always be available. The system vendor could change their system and disable the exporting of documents.

Should that happen (and I'm not saying that it will happen, only that it could happen) then the cloud-based SaaS systems will operate very much like the dedicated word processing systems of the 1970s. They will hold your documents captive and not allow you to export them to other systems. This applies to any SaaS system: word processing, spreadsheets, presentations, emails, calendars, ... you name it. It applies to any SaaS vendor: Google, Apple, IBM, Microsoft, ... you name the vendor.

I'm going to think about that for a while.

The great puzzle of Microsoft Office

What will Microsoft do with Office? Or, what should Microsoft do with Office?

Microsoft built an empire with Office. Office was the most powerful word processor and spreadsheet package. It used proprietary formats. It read files from other word processors and spreadsheets but did not write to those formats, making the trip for data one-way: into Microsoft Office. Through marketing, fierce competition, and the network effect, Microsoft convinced most businesses and most home users to use (and buy) Microsoft Office.

Those were the days.

The world is changing.

Large businesses still use Windows for their desktop environment. Small businesses, especially technology start-ups, are using Mac OS or Linux.

Large businesses still use Microsoft Office. Small businesses are looking at LibreOffice (an open source desktop package with word processing and spreadsheets) or Google Apps (an on-line office package with word processing, spreadsheets, e-mail, calendaring, and other things).

The tablet world is dominated by iOS (on iPads) and Android (on just about everything else). Windows holds a tiny share. The same goes for smart phones.

These are the pieces of the great puzzle that Microsoft must solve. What is a software giant to do?

First, some observations.

Microsoft is the latecomer Microsoft is late to the market, but they have been in this position before and succeeded. They were late with C#/.NET after Java. They were late with Internet Explorer after Netscape Navigator. They were late with spreadsheets after Lotus 1-2-3. They were late with word processors after Wordstar and WordPerfect. They were late with databases after dBase and R:Base. Being a latecomer has not doomed Microsoft yet.

New hardware platforms Microsoft must live (and compete) in a world beyond the PC. Phones and tablets must be part of the solution. Tablets and phones are a very different arena for software design, due to the size of the screen, the touch interface, and intermittent connectivity. Any product on the tablet or phone is a different creature than it's PC counterpart.

Multiple software platforms Microsoft must live (and compete) in a fractured world of software, with multiple operating systems (some not of Microsoft's making or control). Offerings from Microsoft must work with iOS and Android as well as Windows.

The desktop software model doesn't work on mobile devices Microsoft's past technique of selling premium software and obtaining market share through marketing won't work on the mobile platform.

Giving these conditions, Microsoft needs a new approach. Here are some ideas:

Sell services, not software Microsoft will not focus on selling copies of Office for the mobile world. Instead, it will focus on subscribers to its services. The mobile versions of Word and Excel and Outlook will be offered at low prices -- perhaps at no cost -- but they will be useless without the service.

Cloud storage, not local files Microsoft Office will store data in the cloud (Microsoft's cloud).

Not documents and workbooks, but pieces assembled Instead of entire documents and complete spreadsheets, Microsoft services will stitch together fragments of documents and spreadsheets. Think of it as an advanced form of OLE. (Remember OLE and our excitement at embedding a spreadsheet in a document?)

Versioning and tracked changes Microsoft's cloud will keep track of the versions of each document (or document fragment), allowing us to see changes over time and the notes for each change.

Access control (for enterprise users) With all of these fragments floating in the cloud, enterprise users (businesses and their support teams) will want to control access by users.

Promotion and publication (also for enterprise) Users will be able to publish data to other users. Users will also be able to work on new versions of data, reviewing it with other members of their team, revising it, and eventually marking it as "available to everyone". Or maybe "available to selected users".

The idea of Office as a service seems a natural fit for mobile devices. Notice this this vision does not demand Windows tablets -- one can use it with iPads and Android devices. I expect Microsoft to move in this direction.