Hacker News Reader

The Mayor of Munich didn't talk about the feature set available to public employees.

He did talk about making "public knowledge accessible in the future". Public institutions have a moral and legal obligation to preserve certain items for posterity.

The problem with a proprietary application like iWork is that it uses a closed file format that may or may not stand the test of time. I've worked with archivists in government institutions, and they are already struggling to preserve digital documents produced as late as the 1980s.

As a citizen, I have the ability to learn from the primary sources that are hundreds of years old. One of my college friends spent a semester examining the papers of colonial New York State governors and looking at maps and treaties produced with Indian nations.

Unfortunately, the chances are very high that our children will not be able to read many of the important papers produced from the 1980's through today. How accessible will WordStar documents be in 2060? Or Excel XP spreadsheets in 2100?

There are different approaches to improving the current situation. The State of Washington chose to convert everything from native format to TIFF or PDF images. South Africa and cities like Munich are choosing to mandate the use of free and open native formats. There are ups and downs to each approach.

But the free market says that the private sector will save us!

If you throw enough money at the problem, the private sector probably could figure out a way to open ancient WordStar documents. But it'd be so much easier to avoid the problem in the first place...

I'm sure I could given a bit of time write a program to open wordstar documents even without a spec or reference implementation. I imagine the text itself is just stored as ASCII (even if it's not you could just brute force possible encodings until something legible came out). You might not manage to get the formatting 100% but that isn't really any worse than degredation of physical medium.

Hell we cracked the german army and naval codes without even really knowing what a computer was.

I completely agree. I have documents I wrote in the late 1990s I can no longer open easily. As an example of how quickly technology can change this is a perfect example of digital obsolence (storage medium, not software): https://secure.wikimedia.org/wikipedia/en/wiki/BBC_Domesday_...

OpenOffice is far better than recent Microsoft office for opening old Word documents.

He did talk about making "public knowledge accessible in the future".

Yes, but the conclusion does not follow from this goal.

There's no reason that an open format will be readable in the future. If I define some document format, build an application that reads/writes it, and publish it to github with GPL licensing, then in all likelihood it will be forgotten completely in a decade. My inability to gain users will ensure that it withers.

On the other hand, even if the .DOC and .XLS files made by Microsoft are proprietary, they've been reverse-engineered and are well-understood; the .DOCX and .XLSX file formats are even documented. Because they're so universally used, out of necessity, someone in the future will be thinking about how to read those old files.

So being open might make it more likely that something obscure will be readable in the future. But I don't think it's nearly so strong a likelihood as what demand from an army of users of proprietary software will lead to.

Open is just part of the equation. ODF is also an international standard: ISO/IEC 26300:2006 Open Document Format for Office Applications (OpenDocument) v1.0

The point is, in the distant future, someone will be able to use the specification for ODF and read a document that complies with the standard. All aspects of the file will be reproducible, including metadata, tables, etc.

And as I said, open formats are just one approach. The approach that the US Courts and the State of Washington takes (conversion to PDF/A or TIFF) is a valid approach too.

If all they have is the document and the ODF specification, they are out of luck. They'll need the OpenOffice or LibreOffice source code, as that is the real specification.

Both Google and Microsoft were able to independently produce ODF implementations from the spec.

No, they looked at things outside the spec. That's the only way to do it with ODF 1.0, since the spec is massively incomplete.

For instance, it does not specify how formulas work in spreadsheets.

but its in 1.2, right?

Luckily, the source code for LibreOffice is not a secret.

At least someone will be able to store the full specification of the format, or a reference implemetation, along the file that is going to be preserved. It's nothing more, in fact, than one document that refers to another in that the only way to get all the information is to decode both.

You can't store the full spec of a format that's not fully documentes nor keep a runnable reference implementation that has only a binary executable (unless you also preserve the computer it runs on). But then you have turtles all the way down.

x86 emulator shouldn't exactly be impossible to build.

The x86 is the easy part (and, mind you, it's hideously complicated). Then you will have to emulate the rest of the PC - VGA, USB, peripherals...

Isn't it much easier to just have compilable source code?

Let me offer a counterpoint: I started out with Word 3.0 for DOS, in 1987. From there, I went through lots of office software for PC, Mac, and Atari. A good ten years ago, I installed OpenOffice, which I have been using ever since. I never had any .doc compatibility problems.

Today, I'm working for a small retail company in Madrid, Spain: Ten people, running Windows, Linux, and Mac OS. All of us are using OpenOffice. It already was that way when I came here a year ago. Until now, I heard nobody even mentioning our office software. And these people are not "computer geeks" by any measure; we sell clothes.

I guess it is the reliance on Excel macros, which are not compatible with OpenOffice Calc, that prevents lots of companies from switching. If you don't depend on any of those, I honestly see no advantages in using MS Office.

I can only tell you that my mom regularly had issues with people sending her .doc files, which she edited in OpenOffice and they came back somewhat garbled.

All the more reason to avoid the problem altogether and not create unreadable documents in Office.

Seriously: it's not like someone won't be able to buy a copy of LibreOffice. OTOH, a government should avoid mandating the use of proprietary solutions because they exclude people who can't pay for them.

I wish I had your good fortune. I do almost all my development on Linux, and clients would routinely send me Word documents. Most of the time OOo was good enough for reading the content, but formatting was always in some way wonky. I could have lived with that except there was always the occasional document with table formatting OOo could not handle. In these cases the file was essentially unreadable.

I would often fire up Windows, load the office docs I had been sent, and save them out as PDFs so that I could at least have reliably readable files on my development machine.

I had zero confidence that OOo could correctly generate any complex document that would render identically in Office so I either sent clients PDF files created in OOo or authored the documents in Office on Windows. In fact, I much prefer using Word for anything moderately complex simply because document manipulation is just so much easier. (E.g. resizing all text in a document.)

I've seen a number of testimonials from people saying OOo has worked out perfectly fine for them, and I've no doubt it's true for their circumstances. But it does not have the degree of Office interchangeability needed to make it a complete stand-in when you have to exchange Office docs that go beyond simple formatting.

I've not used MS Word for about 6 years but prior to that I always found that OOo (or sometimes KOffice or Abiword) was as good at opening MS Word documents as Word itself.

Even docs specifically created in the format of the target version didn't seem to do very well.

Presumably MS have solved this now and the current versions will produce pixel perfect renderings of all the past versions that LibreOffice can do (and a few more)?

How are MS now at supporting ODT format and importing PDFs?

There's a bigger problem than usability - I say this because each of the places that adopts any kind of workplace software usually has IT/helpdesks/consultants to handhold this kind of stuff.

The biggest problem is really the spreadsheet - people can live with even Google Docs for documents and even presentations. But Excel has managed to be irreplaceable.

When billion dollar termsheets (and all its macros) come bundled in the XLS format, there is no way we can get away with anything less than 100% Excel compatibility. I speak of this from experience in a failed attempt to port an accounting office to Linux - Excel was the only roadblock ... and I tried everything including Softmaker.

I am willing to wager real money that if LibreOffice Calc becomes 100% compatible with XLS 2000 and XLSX formats, we would see the adoption of Linux just explode. Word and Powerpoint dont matter as much - web based offerings like Prezi are increasingly compelling alternatives.

The majority of government computers in Munich already run on Linux and they run well. It does work for them. Munich rocks by the way.

And if anyone's interested - "majority" is ~65%* in _eight_ years into the project:

http://en.wikipedia.org/wiki/LiMux

* 9000 of 14000 as of December 2011. However, "LiMux is a project by the city council of Munich to migrate their software systems, including 14,000", so, probably the total count is actually more than 14000, which makes the percentage even less.

Why is eight years significant?

Because of someone is spending public money on a dubious cause that takes forever to implement?

Considering the number of people currently using an operating system that is over a decade old, I think you don't quite understand the pace at which large, particularly government, projects tend to move.

Eight years for 65% is quite impressive actually.

Regardless, this is quite the tangent...

Let's see: LiMux was based on Debian Etch (2007) until 2011, when it was moved over to Ubuntu 8.10 (Oct 2008); AzLinux, the sole survivor of similar attempts, is based on Suse 11.2 (2009) and only moving towards 11.4 (2011), but it's still in development. For the record, Windows XP SP3 was released in April 2008. So, "a decade old" OS (2008 in fact) is being replaced by basically OS from the same time period over the course of 8 years and is still in progress. I'm sorry, if I don't think it is impressive.

"Windows XP SP3 was released in April 2008."

LOL

You missed out the part where this is ahead of their goal. Instead of wasting money by trying an ambitious big-bang transition with according requirement for additional staff and support, they've purposefully planned out a gradual, and orderly transition.

As for "spending public money on a dubious cause", they're spending roughly as much as before, but they're spending it in the local economy instead of sending it to the US, before you even count any other benefits.

...And of course downvoting for pointing out that 65% of the project in 8 years might be not as rosy, as someone would paint it to be. Ah, gotta love to have a non-popular opinion!

You are being modded down because the 65% is ahead of their original schedule. This means they are investing less public money than originally planned in the migration and spending less in Microsoft licenses than they would be had the project been behind schedule.

I would imagine that employees would get training for the software they use so the benefits of the software being easier to learn on your own isn't an issue for government workers or really any place where they train their employees.