Slashdot Mirror
Charlie Stross: Why Microsoft Word Must Die
Jeremiah Cornelius writes "Rapture of the Nerds co-author Charlie Stross hates Microsoft Word, worse than you do. Best of all, he can articulate the many structural faults of Word that make his loathing both understandable and contagious. 'Steve Jobs approached Bill Gates... to organize the first true WYSIWYG word processor for a personal computer -- ...should it use control codes, or hierarchical style sheets? In the end, the decree went out: Word should implement both formatting paradigms. Even though they're fundamentally incompatible... Word was in fact broken by design, from the outset — and it only got worse from there.' Can Free Software do any better, than to imitate the broken Microsoft model? Does document formatting even matter this much, versus content?"
No sane metamodel. No access from multiple applications. No sane way of creating compound documents. When you see the landscape of modern IT and you notice that the closest thing to that is the XML ecosystem, you know something has gone horribly wrong.
Ezekiel 23:20
somehow, the major publishers have been browbeaten into believing that Word is the sine qua non of document production systems. They have warped and corrupted their production workflow into using Microsoft Word .doc files as their raw substrate, even though this is a file format ill-suited for editorial or typesetting chores. And they expect me to integrate myself into a Word-centric workflow, even though it's an inappropriate, damaging, and laborious tool for the job.
So his publisher is forcing him to use Word. I would be annoyed as well. I know at least some publishers accept PDF (and some even LaTeX). So maybe he should just choose a different publisher.
"First they came for the slanderers and i said nothing."
It's called Lyx - http://www.lyx.org/
Yes, let's enumerate all the structural untidinesses of Word. Let's blame that application -- which held its own, against many, many competitors, not because of a megacorporation strong-arming it (remember, MS was not always a megacorp) but because it was good at doing what users wanted it to do -- for the inelegance of its data model. Let's compare it to SGML, which is so much nicer and easier and so much more elegant if you're a programmer and can appreciate that sort of elegance, and if you're not a programmer, well then for god's sake why are you touching a computer?
If you want SGML, you know where to download it.
The opinions stated herein do not necessarily represent those of anybody at all. Deal with it.
This critic just comes across as whiny to me. I use Microsoft Word to typeset complex multilingual documents, and it works great for my needs. I've occasionally tried to use Scribus and some other OSS tools, and have been blocked by limitations, typically related to non-Latin text handling. Word is also very scriptable from pretty much any programming language via the ActiveX interfaces, which is how I use it.
If he has a better idea of how to set up a word processor, he didn't see fit to share his thoughts with the rest of us. But serious suggestions only, please. If the author wants Microsoft to make Word more like vi, I think then we'd really see some "loathing both understandable and contagious" from ordinary users.
TeX is horrible in the sense that it's a glorified macro processor state machine, with every character in the input stream potentially changing everything in the output DVI - text itself, formatting registers - you can't tell form the character other than by continuing processing the Turing-complete macros, all of them capable of changing every aspect of the processor state! Change that one character? Regenerate the whole document! Good bye, interactivity. Good bye, inspectable structure. It's perfect for the hard-core typesetting jobs, but lousy as a document platform in any other sense than the typesetting one. (But it would be nice to have good exporters for it in as many applications as you can get. And when I say "good exporters", I mean template-driven export engines with fine-grained tuning, not some "just export it, I'll tweak the styles by hand" thingy. Remind me, why do we have computers again? To automate stuff? I thought so.)
Ezekiel 23:20
Joel Spolsky has an excellent write up on why the Office file formats suck. A must read.
http://www.joelonsoftware.com/items/2008/02/19.html
He actually worked on Excel leading to funny anecdotes like this one
http://www.joelonsoftware.com/items/2006/06/16.html
... and since discovering it, I have felt no need to use MS Word for anything anymore. Particularly good about LibreOffice Write? The PDF export function works flawlessly, exports quickly, and also gives control over how the PDF document appears in Acrobat Reader (Zoom level, page order, thumbnails, et cetera). To me, Word has had its day. LibreOffice Write works well, is free, requires no internet-licensing shenanigans and does everything one could expect from a good word processor, including auto spell-checking and thesaurus functionality. My 2 Cents. =)
Why did the chicken cross the road? Because Elon Musk put an AI chip in its head.
Yes, I still pine for WordPerfect 5.1, and even the early Windows versions.
Three words: Reveal, Codes, and Acerson.
With just those you could do damned near anything.
To this day, likely close to ten years since I stopped using WordPerfect, I still find myself clobbered by strange MS Word formatting edicts, with no obvious way to get rid of them.
At least with WP you could see why something was weird, and fix it.
Three Squirrels
I've used Wordstar, Wordstar 2000 (or 3000?), WordPerfect, MS Word, and OpenOffice/LibreOffice writer and they all pretty much suck. Most people misuse them. They don't integrate well with other software. And they produce ugly results.
I wrote my master's thesis using FrameMaker which was quite a bit better. However, for my current document-production needs, I use LaTeX. I maintain the manuals for my company's software products and we have a great workflow for building the manuals. The same Makefile that builds the software also builds the manuals: PDF versions directly from the LaTeX and HTML versions using htlatex run on the LaTeX sources. Then a post-processor fixes things up so that our HTML documentation is linked context-sensitively from the web pages of our app, and special goodies like embedded training videos are placed in the HTML documentation at the right place.
The power and control we get from this workflow is unmatched.
That about says it. Nobody else cares. I've been using Word since it came on two 5-1/4" floppy disks and included a mouse and used every version since what? 1983 or so? (Before that I used Zardax on an Apple ][ and, of course, WordStar.)
There's not a damn thing wrong with Microsoft Word. It is quite adequate--superb, even--for 99% of the people 99% of the time. I've written several 300 page books with it, including extensive indices, sidebars, tables, graphs, and pics and it works just fine. No, you can't do EVERYTHING you might want to do with it. And you might actually have to put some time in learning how it works, but ONE thing is CERTAIN:
It's not going to go away. The chances of it going away are equivalent to the chances the United States will convert to driving on the left. Only the nerds care about the arcane details under the hood.
Nobody else gives a rip.
How about a moderation of -1 pedantic.
enough said!
MS Office is designed for use by the 9-to-5 clerical worker --- not the outside studio or in-house team that designs your four color catalogs, print adds, brochures and annual reports.
My number one wish for any word processor, but especially Word, is a switch that says:
I'm writing a document that will be printed out on paper with black toner.
At a minimum, I don't want e-mail addresses or URLs changed to blue, or underlined, or hyperlinked.
My number two wish is a switch that says:
Anything pasted into this document will adopt the formatting of the line into which it is being pasted.
I cannot think of a single instance, ever, when I wanted the formatting from some web page to be carried over into my document. My final wish is to find a word processor that assumes, or at least makes really easy to specify, that the Page One Header will not be used on subsequent pages. I don't recall how Word does that these days, but in LibreOffice it involves creating a style just for the first page. Assuming that you've managed to Google the specific forum post that tells you that.
Three Squirrels
Gee, I can't imagine why people would use Word over this.
Christ. This article, like so many here at Slashdot, summarizes to: Usability matters. Usability matters A LOT. Open source developers still don't fucking get it.
Here's a thought: if you want people to stop using Word, why not make something better than Word? Shocking.
Comment of the year
Of course, there's the TeX engine and the default macro packages which are different things.
Actually, I was talking specifically about the generic TeX engine properties that make it virtually impossible to use it for incremental redisplay and similar purposes, so you have to go WYSIWYM instead, but then TeX isn't the foundation for the document platform itself but just the output module, which is what I was pointing out. It also doesn't work well for compound documents, not just simple hand-edited text ones ("simple" as in "not compound", but not excluding structured and large ones, of course), but that's sort of outside of TeX's scope (if you ignore Web2C's [\immediate]\write18, or yet another converter to "lower" the document data (for example, data linked to a live online resource) into something "set in stone" that TeX can actually typeset), so perhaps that would be an unreasonable request anyway.
You are not bound to plain TeX or LaTeX or a few others, write your own.
That doesn't solve the problem I was referring to. One potential solution that occurred to me the other day was that since TeX's state tends to be small, and the processing is "paged" in the last stages of TeX's internals, one possible option would be memoizing the internal state of the processor and resuming the execution only from the nearest place preceding your edit point in the input stream where the last state snapshot took place and going on from there. But that still requires a modified TeX implementation, and it doesn't work for multi-pass processing that, i.e., collects page number references to scatter them throughout your document in a later pass (a single inserted character could change the page number on a page reference on the very page you're editing right now, and there's no way to know without running all the full document passes again, and there goes your real-time redisplay), and that tends to be quite common, so tough luck.
Apparently you need to be reminded. TeX is a system for "automating stuff", not a system for interactively enabling tweaks by hand.
Apparently, I failed to get my point across. Of course TeX can do that, but the document application with TeX export has to export the document structure into the TeX file properly, and I have yet to see a complex application doing it in a user-friendly way. Nowhere am I suggesting that *TeX* is the right place for the necessary user interaction. If you're accusing me of a desire to do stupid things, you had better point out which ones, since I fail to see which part of my requirements is unreasonable.
Ezekiel 23:20
So much this!
Word is popular because it is can be manhandled by just about anyone with basic computer skills into producing a passable document. Most people doing anything more complex have a variety of tools available with various trade-offs in functionality and learning curve.
This is largely true about most popular products. Tools which are technically inferior for a purpose are used because the gap in technical fitness is smaller and lest costly than the difference in skill required to use the better suited tool. Cheaper to have the hammer than can pound in most nails when you don't really care as long as it holds, then go get the specialized hammer when you do..
Do you really believe Word has advanced and improved since 2006?
There is a recurring problem with software development in recent years, where one player has become dominant, the barriers to competition are so high that it has no real challenger for a long time, and the result is stagnation. There are numerous examples: Microsoft Office for business documents and spreadsheets, Adobe Creative Suite for graphics, Autodesk applications in the 3D modelling space, IE6 as the classic beloved of web developers everywhere, and as an odd one out just to make the point, Linux if you want an OSS operating system.
There are a few ways out of the trap, but the big problem is that the people making purchasing decisions often aren't interested in assessing the quality or productivity benefits of alternative software, or even able to make informed judgements about those things if they wanted to. No-one ever got fired for buying the market leader, so while they know that the new subscription pricing model will give vendors even less incentive to actually improve anything or the support contracts are probably far more expensive than they're actually worth or the TCO will be horrendous because of usability problems, they'll carry on using these leading products anyway so their careers aren't at risk.
That creates a vicious circle where no-one is willing to invest the staggering amounts of time and money required to build a heavyweight competitor that can effectively challenge an incumbent. Instead, we get open source clones or cheap-and-simple web/mobile apps, which do a good enough job to save some users paying for the heavyweight commercial software, but in most cases offer little real innovation and almost invariably lack the quality and feature set of the established big names. That's why the professionals spending serious money keep buying those big names, and so the cycle continues, with little incentive for software giants like Microsoft to improve their cash cows or innovate with entirely new products.
I think the most likely way out of this in the long term is for a new product to arrive that changes the rules and moves the market. With formal printed documents becoming less popular and an increasing emphasis on on-screen presentation and collaborative editing, is a word processor still a good model to manage business information? We have far more powerful (and systematic) formatting capabilities in numerous browsers that can render HTML+CSS content. Probably every programmer reading this routinely uses far more powerful editing, review and collaboration capabilities in their everyday tools. I don't just want Word 2014 any more. I want something that helps me collect, organise and share information in ways that match how we'll be living and working in 2014. And a tool that does that might have a small chance of breaking the Word stranglehold.
If you disagree, post your argument. (-1, Overrated) isn't your personal censorship tool for views you don't like.
What exactly makes Word so bad? It seems functional enough, and I fully admit that maybe I'm just not understanding the finer points of some programming strategies, so what's the deal?
Let's see, just a few off the top of my head:
- Terrible flow control. It you change page one, have fun tweaking all the rest of the pages to get things to line up.
- Lack of frame control. In order to create a large document or book with complicated multipage graphs or graphics, you need a strong set of rules for where to break up rows, etc. Not to mention the flow of any text around the frame.
- Non-organic styles. There is no easy way to change the style of logical parts of the document globally, for example, change the size of the font on the headers of a certain class of tables over all chapters. Or to put it another way, global definitions plus exceptions.
Word users are just used to the constant re-treaking of pages to make them look right. Just another example of Microsoft office leading to massive wastage of man hours.
The only thing worse than a Democrat is a Republican.
you are supposed to write things. Then you twiddle how it looks a bit at the end.
That's hardly how any business document in existence is written. Layout is considered at the same time as the content. Presentation is often more than 50% of the value of the document. It is essential to be able to make edits right in the final output. Nobody is willing to print the DVI, then mark it up with a red pen, and then to find corresponding code that programs that piece of the document, and then to change it ... and once you change something on page 1, things cascade down to page 100 - pictures and tables jump onto different pages, blank areas show up where none were before... this means you need to redo the DVI and review after every change. To compare, a WYSIWYG wordprocessor gives you exactly what you are going to print for given page settings. You just go from the first page down and make your changes. That's why MS Office (and WordPerfect before that) rule the office space, not TeX. Those wordprocessors do a pretty good job on having things done your way, with background pagination and other niceties.
They're all solving different problems. TeX is designed for typesetting, which HTML and Word formats aren't well-suited. HTML does structure well, but it's useless for typesetting. OOXML is a weird mix, not really well suited to either task. It's better than older formats, but it's still incredibly painful to generate, and near impossible to read. If I had to guess, it's designed to give MS the ability to say that the format is open, while still making it difficult for competitors to support.
Required reading for internet skeptics
I have crappy memory. I'm not one to remember arcane macros or shortcuts. Ctrl-C and Ctrl-V is as far as I go.
Never fear - nested menus always helped me find things.
Now with the Ribbon in Office, first I have to search each of 5 or 6 ribbon views. They aren't grouped very meaningfully so it's basically a linear search. Then if I don't find what I am looking for I am basically stuck, since I can't remember how to find items that aren't on the ribbon, and I can't really search on them, because I am not exactly sure what they are called. So I end up using Word like a glorified notepad.
The design team that killed the menus on Word (and those kids who are doing the same on browsers) don't realize the damage they are causing end users.
The primary argument used to dismiss Libre Office as a viable alternative to MS Office is that it "can't do anything more complex than basic document editing".
And yet here we are, explaining why MS Office is the popular choice because it only does the most basic stuff that 9-to-5 clerical workers need.