A National Archive Moves to ODF

Andy Updegrove writes "The National Archives of Australia (NAA) has announced that it will move its digital archives program to OpenOffice 2.0, an open source implementation of ODF. Unlike Massachusetts or the City of Bristol (which announced it would convert to save on total cost of ownership), the NAA will deal almost exclusively with documents created elsewhere in multiple formats. As a result, it provides a "worst possible case" for testing the practicality of using ODF in a still largely non-ODF world. If successful, the NAA example would therefore demonstrate that the use of ODF is reasonable and feasible in more normal situations, where the percentage of documentation that is created and used internally is much larger."

Comments

Beginning of the Revolution!

[Phantombrain]

I'm wondering if this will be the start of the use of Open Source in more business applications. Most companies use M$ Office, since it is mainstream, even with it's large cost. Maybe the Government's example will be the beginning of the revolution.

Re:Beginning of the Revolution!

[ozmanjusri]

I'm not sure why an office format would be the best thing to use for archives of final documents;

ODF is an electronic document format, not an "office" format, whatever that means. Its advantage in this context is that any document in ODF can be dissasembled ito its component parts easily. Text, images and formatting can all be extracted and used separately if needed. PDFs are hard to convert back to the raw data.

Re:Beginning of the Revolution!

[LardBrattish]

A couple of years ago I went on a tour of the West Australian archives & they said that computer generated documents were their bane.
They had 150 year old documents going back to wherever but they had trouble reading 25 year old floppy disks in weird formats and converting them to the raw text-only format they used back then.
If they standardize on an XML based format like the ODF ones and convert all of their old stuff to this it will make archiving the current documents much easier. It may even in a few years prod the Australian government to standardise on a product that saves to ODF...

Worst Possible Case?

[multiOSfreak]

As a result, it provides a "worst possible case" for testing the practicality of using ODF in a still largely non-ODF world.

Wouldn't this sort of test be a more or less good test case for switching to ODF and dealing with non-ODF outside documents? Maybe I just misunderstood the comment.

Re:Worst Possible Case?

[qwijibo]

That's the point. The real world plethora of formats is the worst case. If ODF can handle the worst case, it would be a testament to the robustness of the format. The worst case test for interchangeable file formats would demonstrate that ODF is viable.

Re:Worst Possible Case?

[MrPower]

What I think they meant to convey is that this will be a worse case scenario they can use for testing the practicality of using ODF in a non-ODF world.

But I don't actually think so...

Whereas I think this will be great for ODF, as the NAA will have to produce heaps conversion software to convert many formats to ODF but because they are an archiving operation, they won't ever have to convert back. Instead, I imagine that the common document format for outgoing files from of the archive will most likely be PDF...

This scenario won't test the ability for ODF in collaborative work among entities, something that I would see as the worst case scenario needed to test the practicality of using this format.

Having said all of that - to hell with everyone else - I have been using non Microsoft formats (first Star Office formats and now ODF) for five years now and rarely come across a problem. Then again, I am a simple user so I wouldn't expect too much grief. From my experience advising other people I can see that the true hurdle is not the file format, rather the application. Word and Excel are automated from so much business and scientific software that people just expect the results of their query or analysis to be dumped directly into their spreadsheet or word processor. So until Quicken or MYOB support something other that MS software, or until alternative software is produced that does, business will largely use MS.

On the other hand I strongly recommend to people to use OOo at home and with the ever increaseing compatability that OOo has with MS formats, this is not a bad option.

Novell has great success with this

[Sir_Jordan]

Years ago when Novell switched over to Linux operating systems, one of their largest fears was the trouble integrating their documents in a Microsoft stardard based world. It turns out that Open Office was more than adequate concerning reading/writing various document standards.

Doesn't exist?

[Anonymous Coward]

Which of these applications, exactly, don't exist?

OOo shows how bad Java can be

[jimmyhat3939]

OOo is slow because it's still largely impelemented using a Java VM-based architecture with bytecode and all that entails. I really think these guys should reconsider. MS is moving toward an XML-based file format which shouls be open enough for anyone. And MS Office is a client app written completely in optimized Windows assembler code. That should help with performance hemi-dramatically.

Re:OOo shows how bad Java can be

[olafura]

You must be kidding OpenOffice.org is almost all C++ code, it's slow to start because it calls to many files on startup, it has certain parts of it which use java like the Base and some templates.

Re:OOo shows how bad Java can be

[LWATCDR]

"OOo is slow because it's still largely impelemented using a Java VM-based architecture with bytecode and all that entails."
No it isn't. I just ran OpenOffice writer V2.0 and checked my task list. No java was running at all!
OOo uses java for some functions but it in not "largely impelemented using a Java VM-based" anything
http://en.wikipedia.org/wiki/OpenOffice.org#Java_c ontroversy is a list OpenOffices use of Java.
OpenOffice is mostly a C++ or C program.
I have not run a profiler on OOo so I can not tell you 100% what makes OO slower than Office but I would guess that part of it is the XML format that OO uses.
Just from my own experence I have found that you can write a fast XML parser and you can write a "safe" XML parser. But a fast safe XMP parser is very hard.

Re:OOo shows how bad Java can be

[iabervon]

The slowness of the XML parser doesn't matter much if you don't save and reload the document with every modification.

Actually, OOo is so slow because they don't use a widget set. The display is hand-drawn by a bunch of monks in Germany, because the project started before Qt or Gtk.

Re:OOo shows how bad Java can be

[the_womble]

Turning off java does speed up Openoffice considerably.

So does increasing the memory settings.

However it still takes about 3 or 4 seconds to start up on my desktop. As far as I remember from when I still used Windows this is not all that different from MS Office on XP on similar hardware. Does any one else who has done the same tweaks differ?

However Abiword or Lyx starts instantly. I mostly use Lyx (which I find more productive) and Gnumeric (faster, with some nice features) rather than OO.

Re:OpenOffice.org 2.0?

[99BottlesOfBeerInMyF]

Note: Said software doesn't exist.

Get with the times. That hasn't been true for a while. The current list includes: Abiword 2.4, eZ publish, IBM Workplace Documents 2.6+, KWord 1.4+, NeoOffice 1.2 Writer, OpenOffice.org Writer, Scribus 1.2.2+ , StarOffice 8 Writer, TEA text editor , TextMaker 2005, Visioo Writer 0.6, and Writely for the word processor portion of the format, with similar lists for the other components. There are a lot more that have announced support on the way.

Re:OpenOffice.org 2.0?

[Distinguished Hero]

Wikipedia begs to differ.

Some highlights according to wikipedia:

.odt: AbiWord, KWord, Writely
.ods: KSpread, Gnumeric (incomplete)
.odp: KPresenter

Plus StarOffice (maybe that's cheating), and IBM Workplace Documents (never used it)

Re:OpenOffice.org 2.0?

[R.Mo_Robert]

KOffice?

Not that that makes it a myriad, but there are also a few lesser-known programs that do, and I would guess that many others will implement support for it soon. AbiWord didn't last time I checked, but they did support SXW (StarOffice/OpenOffice.org Writer 1.x format), so it wouldn't surprise me to see them implement ODT. Actually ... oops ... I lied, looks like it does now: http://en.wikipedia.org/wiki/List_of_applications_ supporting_OpenDocument

Anyway, the OpenDocument Alliance also has a lot of companies behind it, among them IBM and, of course, Google. So it seems to be a pretty strong format to me, even if that one company from Redmond (what's their name again?) isn't particularly interested right now...

Re:OpenOffice.org 2.0?

[Distinguished Hero]

One should not copy and paste without citing the source. List of applications supporting OpenDocument seems to be where you got that form.

Re:First

[Luctius]

I think the dutch national archive also switched to odf.

Re:Bristol, UK?

[Fluoxetine Freak]

http://news.zdnet.co.uk/software/applications/0,39 020384,39260274,00.htm

Yes, Bristol UK.

Re:Bristol, UK?

[caluml]

Hmmm - interesting. Wonder if my Council Tax bill will go down in line with any savings they make?

Re:Bristol, UK?

[Anonymous Coward]

Bristol's population is about 400,000. Bristol aims to save £1,000,000 over five years, or 50p per person per year. Only one fifth of council revenue comes from council tax, so your bill should be reduced by about 10p. Don't spend it all at once!

Small experience

[Anne Honime]

Back in the Uni, I was in charge of merging some 20+ articles from various authors into a single document. The target was to give the publisher a uniform document which he would then transform into a book.

All documents were made with a flavour of Word or another, from word for MacOS 6.0 to the latest (at the time) word XP for windows. As you'd have already guessed, the only word processor able to make sense of all the documents at once was Openoffice.org. Of course, I faced issues (bulleting appearing "funny", for instance), but as I was applying a style I created, that was not a problem as long as the text was there.

No single version of word in my possession was able to open all the documents, some documents even crashing word XP with thunder and lighting.

Questions here

[countach]

I wrote the original version of the National Archives software that does the conversion. The current version of the software is available here: http://sourceforge.net/projects/xena

If anybody wants to ask any questions here I'll try and answer.

Taken

[tepples]

OK why is the little o included in the name? Its just Open Office. OOo is a website that Has OO. I don't get it.

If this Wikipedia article is to be believed, then the name of the web site, project, and product is "OpenOffice.org" because "OpenOffice" was taken.

Re:OpenOffice.org 2.0?

[Noksagt]

which apps on the list 1. run on Microsoft Windows operating systems (so that they don't require re-buying hardware)

Those which run under Linux probably wouldn't require new hardware either. That being said, here are the windows apps (which are most of them):
ODT

• Abiword
• EZ publish
• IBM Workplace
• Scribus
• TextMaker

Writely is web-native, so you could use that too.Kword might work in cygwin (I really don't know--I know you can run some KDE apps).
ODS

• Gnumeric
• IBM Workplace

Same note on KSpread.
ODP
Same note on KPresenter as on KWord
ODG

• Scribus

2. are promoted in print or on television across North America or across Europe?

What does this have to do with anything? I have seen relatively few MS Office, OO.o, or Corel WordPerfect ads either. People giving away software usually don't spend money to ensure you'll take it from them.

Re: templates

[michaelbuddy]

I almost thought you were joking about the templates, because what you described is pretty exactly what some people have done. It's called OOextras.

I don't think they match up to the beauty of (some) MS or Corel templates , but StarOffice has some templates you could steal from I bet. Would those be freely distributable under their license?

Anyway, http://ooextras.sourceforge.net/

that's the

Driver issues; marketing

[tepples]

Those which run under Linux probably wouldn't require new hardware either.

Find me a Linux driver for my paid-for yet unsupported Microtek Scanmaker 4850 flatbed scanner, which was purchased long before I thought of switching this computer to Linux, and I'll believe you. Unless you are working with a computer that was built from the ground up for Linux, including buying a printed copy of a distribution's hardware compatibility list to carry with you to the computer store, I am 90 percent sure that you will have issues with at least one piece of hardware if you switch a computer from Windows XP to a common Linux distribution.

And what about vertical-market proprietary software intended to run on the same computer, which is either available only for Windows or (if you're lucky) available for multiple platforms but priced such that using multiple platform versions in an organization is cost prohibitive? You would have to use Wine (significant overhead and less than full compatibility) to run your existing licensed software for Windows on a Linux box.

What does [promotion in traditional media read by management] have to do with anything?

It's the same reason most listeners prefer payola'd major label music to independent music: repeated exposure builds familiarity.

I have seen relatively few MS Office, OO.o, or Corel WordPerfect ads either.

Which magazines and which TV channels are you looking at? In the news magazines and cable news channels, I see a whole bunch of advertisements for Microsoft Office software.

People giving away software usually don't spend money to ensure you'll take it from them.

Then why doesn't Sun advertise its StarOffice software, the official commercial distribution of OpenOffice.org? Or by "giving away software" do you also mean "we're practically giving it away", that is, budget software?

OOo2.0 is just one facet

[digipres]

Our use of the OpenDocument format will be quite important, but it's only one facet of what we do. The Xena software has been developed with a plugin architecture that lets us use various external helpers to 'normalise' or convert to open formats any data objects in our care. For each data object, we use Xena to create a base64 encoded copy so that we can embed some metadata with it, and separately for a conversion to an open format. Much of the data ends up as XML, while images for example are png or jpg. We're currently investigating open audio formats. Xena is also used to 'present' data objects that it normalises.

Until now, Xena has made use of OOo 1.1.x for the normalising of office documents into flat XML. Other development priorities have kept the move to OOo2 in the background. I must stress that we have not yet released Xena with OOo2 support, there is more testing to be done and we feel that the release must be accompanied by good user and developer documentation.

The 'current' binary of Xena available at sourceforge is waaaaay out of date and will shortly be replaced by a much sleeker and more intuitive version. For the curious, anonymous cvs is pretty up to date. If you have a java 1.5 sdk and apache ant, check out a pile of modules and go nuts. Anyone who wishes to become involved in the development effort is more than welcome.

For anyone else, keep an eye on the http//xena.sourceforge.net/ for the upcoming binary release.

Re:Bingo!

[shaitand]

Correction. "OOo is slow" AND "it's still largely impelemented using" C and "C++ with all that entails."

There is certainly no reason to believe it is slow BECAUSE C++ was used. One can write a slow app in any language. It is just a bit easier to do in an interpreted language like Java than in a compiled language like C++.

P.S. Don't tell me that Java compiles to bytecode. That just means that Java compiles to an interpreted language instead of a native language.