But once again we see this described as "exporting." The focus is on the silo. You're in the silo or you're not. Rather, the focus should be on the data, and applications should merely manipulate them.[^1]
[^1]: In many ways, there is a similar distinction between programming models. For example, in an "object-oriented" style the object owns the data and the functions on that data. In a "functional" style, the data exists, and there are functions on that data. The analogy isn't perfect, but at least it sheds new light on the word processing issue.
Rich-text, WYSIWYG editors give people more of a feeling of control and perceived ability to see the big picture. Sure, we can argue that this impression is largely superficial and adds more distraction than anything. But that's because we're experienced computer users who know that "code" directly relates to what WYSIWYG displays, i.e. WYSIWYG is just a middleman. But this is something that has to be experienced to really be grokked. If it weren't for sites like CSS Zen Garden that made it clear how style could be completely separate from content, I'd probably still be making pages in Dreamweaver today.
The rich-text, WYSIWYG interface shows the data people want to interact with; they care about the text in their document and how it is formatted on the page, not how it is stored on disk.
So I disagree with z1mm32m4n about the power of plain text. Plain text is useful in Unix because as a storage format it corresponds to the usage pattern of typing text into a terminal; meaning interactive data entry and reading from a file don't have to be different.
But plain text is a lie on all modern Unix systems, it has been replaced by a mix of text and escape codes. Recently I tried to cat initial-commands.txt - | some-interactive-program and did not get the expected result because the pipe between cat and the program was not a tty. (The solution was to use script to enforce tty-like behavior.) The reason these escape codes exist is that people wanted more than plain text could give them (e.g. color).
However, this is not to say that a universal storage format for data of a kind (e.g. formatted text) would not be useful to avoid siloed applications; if you are working with data that already has a standard format, inventing your own and requiring import and export for everything else is just silly.
At school, pupils are taught Microsoft Word most often, and LibreOffice at best. They know nothing but WYSIWYG and separated applications that don't talk to each other (Microsoft Office can be scripted with VBA, but they don't know that).
The fundamentals of computing are not taught. File formats are not taught beyond "this app opens that format". Plain text is that ugly stuff you read with Notepad. The internet is accessed with a web browser. And so on.
---
This needs to change. Computing education should be expanded, and other curriculum should be cut or adapted accordingly:
Data formats should be taught. HTML could serve as a first example: you can view it online, you can save the page and view it offline, you can view the source with ctrl+U, or even modify the source with Notepad (Kudos if your school uses a free operating system, but let's not require it right away).
Rudiments of programming should be taught. Something like Human Resource Machine (a coding game in pseudo-assembly) could serve as an introduction βor we could stop there, that game is pretty complete. Now that's quite a heavy topic (I expect Human Resource Machine alone would take a whole semester), so I propose we use the Math class for this, and cut on other math topics (I'm not sure which, though).
The internet and the web should be taught (and not just used). I believe the English class is most suited: the internet is mostly about reading and writing (YouTube notwithstanding). We should cut down on classical literature and literary analysis, and focus more on information processing (reading more or less reliable stuff and come up with reliable information), and public writing. If there's a class that ostensibly teaches citizenship, it should be used to teach public debate, which are arguably more important than the arcanes of your country's constitution.
i just think that kids spend enough of their time on regimented learning. nine hours a day plus homework. we should stop inflating the curriculum and leave some of their mental energy in reserve so they can spend it on self-driven freeform learning and hobbies.
I do not propose we teach actual programming to kids. Just the fundamentals of computational thinking, for which an introduction to programming may work really well, and I expect 40 hours will be more than enough. Even a couple hours may be enough.
Computers pervade our lives. They're becoming the medium through which we do everything. We'd better learn to control them before they control us (I'm thinking of golden cages like the iStuff, and surveillance/advertising machines like gMail and Facebook).
The printing press allowed the people to read. The internet is allowing the people to write. Such a fundamental change in communication structures is bound to have similarly profound effects. But first, people need to learn to write. School isn't the place for that. So far, my best writing school has been public internet forums. School only taught me the very basics, then went on having me read novels I didn't enjoy. Before 1995, this was fine. Less so now.
https://reddragdiva.tumblr.com/post/127964136013/it-turns-ou...
https://reddragdiva.tumblr.com/post/137374628373/scratch-is-...
https://reddragdiva.tumblr.com/post/137376311558/scratch-is-...
Apparently it's not just the academically bright kids, the middling ones get it catching their attention too.
Later they get taught "textual languages", which I'm pretty sure will be Python.
They also get taught to use the office apps - but they are introduced early to the concept that these machines that civilisation runs on are things they could tell what to do.
This is exactly why walled gardens are terrible. It's killing innovation and actual improvement in this case.
There's a growing class of editors that strive to provide a richer document experience while still relying on plain text as the file format. I don't usually suggest normal people adopt it, but an example of this kind of hybrid approach is Orgmode in emacs. I can do ALL SORTS OF THINGS with my org files, but the actual data is still just plaintext with minimal and human-readable markup. The editor does the rest of the work.
I find that promising.
There was an (astute, I think) article from "Joel on Software" years ago saying that the main reason that supporting the Word format was hard was that, by implication, you needed to implement every feature Word had to do it right.
a plain-text light-markup file -- easily read, parsed, and rewritten when necessary by a plethora of apps (text-editors, spell-checkers, grammar-checkers, link-checkers, sanitizers, validators, on and on) -- is a big advantage (gargantuan!) over a complex xml-format document stored in a .zip container.
and editing and ease-of-use (and re-use and repurposability) is much friendlier as well.
light-markup will never do everything ms-word does.
nor should it.
but typical light-markup systems these days handle all of the features needed for the vast majority of long-form documents which are now being produced -- books, annual reports, scientific articles, etc. -- and do it with an eye toward mounting on the web.
whereas ms-word is still stuck in a page mentality, and creates .html output which is rather horrendous.
but i agree that ms-word is still valuable to many.
and i no longer feel it necessary to bludgeon word for being the "de facto default" out in the world, where everybody uses it because everybody uses it, and nobody can use anything else because "conform".
so if someone needs one of the "does more" things where ms-word does more, i'm glad they have word. it's a tool in their chest, no reason to begrudge. i'm happy, even, that word has true fans like you.
i'm also glad the rest of us can move on now to something that we think is better. it's about time.
But you can get pretty far down the road. That's my point.
Again, Orgmode is a great example. I can (and sometimes do) use my org files outside of emacs.
ms-word can support those things as well, but only when you bring its proprietary format into its proprietary, closed-source app, which is not free of cost. the silo only looks better when you are locked inside it.
it's fine to prefer ms-word. prefer whatever you like. but think clearly and fairly if you start to bring up objective comparative advantage.
and i like the simplicity and flexibility and freedom of plain-text as my format of choice. but if you or anybody else prefers something different, i urge you to take that path.
1. Stickiness.
Google Docs (and even Microsoft Word) work so well when you use them in conjunction with their respective ecosystem. Many users choose Google Docs over Dropbox Paper simply because they already have Google Apps accounts, even though as a product what they really want is something more like Dropbox Paper.
Inevitably, what happens is that those who can try out a new product on a small team do so, but those who are forced to work between teams or even between companies choose the solution they're already using; i.e., the stickiest.
2. Plain text is powerful
I'm sure I'm preaching to the choir here, but each new "office productivity" or "word processor replacement" tool comes with it's own silo. Even Microsoft Word, which has it's own file type that you can save to your Desktop, is largely a proprietary format. You can export data out of Google Docs or Dropbox Paper, etc., but the idea is that you're leaving the promised land. You're packing up and moving somewhere else.
On the other hand, we know from the Unix philosophy that plain text is a universal interface. Using plain text can help us escape these siloes. The problem, of course, is that from grade school we teach students that to edit text we use Microsoft Word.
There are a lot of cool new products in this design space of "word processor replacements," but few which tackle these two fundamental issues.