Archive for the ‘OOo’ Category

A simple thought on the management of software

Sunday, June 29th, 2008

I just wrote a column for Charles Arthur about how I had celebrated Firefox day by switching back to Opera and this thought didn’t quite make it in.

The organisational division between the users and producers of software should follow its patterns of use. Open source software is predicated on the idea that users ought to become producers too. Most of the time this is simply false, and when it is, commercial or at least closed source software will work better. The clear organisational distinction between producers and consumers corresponds to the reality of function. That is why Opera is better than Firefox. With the great open source success stories — LAMP, essentially — there has been a more permeable barrier between users and producers precisely because the user/customers are themselves producers of software to start with. Even then, the successful projects are run as despotisms.

The half-failures like OpenOffice have failed to understand which side of the barrier they are on, and to organise themselves accordingly. Actually I am inclined to think that OOo ought to be open source. There’s certainly no commercial justification for another disk-based office suite. But in that case the distinction between Sun as the inner oligarchy of producers, and everyone else as more or less favoured consumers simply won’t hold. The balance, difficult and unpleasant though it clearly is for Sun to understand, is not between professional engineers and happy, laughing amateurs with their intoxicating natural rhythms, but between the Sun-salaried workforce and their competitors from the real world, at IBM and the various Linux companies. The end-users are wholly irrelevant.

Instant bibliographies

Monday, March 31st, 2008

I hate bibliographies, and I suspect that any tool I bought to handle them would consume immense quantities of otherwise productive slightly less unproductive time. On the other hand, I need to have reliable references and footnotes from time to time and I think I have discovered a free and fairly reliable way to get them into Openoffice that requires no more than half a dozen mouse clicks from the moment I find a quote in the book I am reading. It will, of course, work with Word, too.

I get the bibliographic details of whatever I am quoting from Librarything: this requires at most half a dozen passes with the cuecat scanner and often only one. From Librarything, they are dumped into the Zotero firefox extension. That’s one mouse click. Zotero has plugin toolbars for both word and openoffice, so when I need to source a quote, I add one more mouseclick in OOo, and then type in the page number. Instant footnote.

I think it will also produce a bibliography at the end, but I haven’t tested that yet.

Making OpenOffice tolerable

Wednesday, January 2nd, 2008

One of the incidental irritations of using OpenOffice is that it uses its own, eccentric python installation. OOo is almost completely scriptable in Python, but the mechanism was put together many years ago by one programmer in his spare time, and still has some large rough edges. In particular, it doesn’t know anything about system Python libraries or search paths, which is a real pain since half "the joy of the language":http://xkcd.com/353/ for lazy people is that someone else has usually written whatever you want first.

In my case, this means that someone else had written a module which can be used to detect non-ascii characters and replace them with html character entities; another one to make native dialogue boxes for OOo; and several to interact with Movable Type and upload stuff there. So all I had to do to make it as easy to publish to my blog as to print was to wodge all these together and swear at them until they worked.

There was still at least one important snag: that OOo itself is so slow to load that it wasn’t worth using it for a blog editor unless I were in the middle of writing something else, and in that case, I shouldn’t have allowed myself to be distracted.

So whenever I last upgraded my copy of OOo, I didn’t bother copying all the bits for the script into the new directory, and for some time have been doing all this by hand. Then I finally got around to trying "the Novell-built version of the program,":http://go-oo.org/ and it loads in about a quarter of the time of the official Sun build (that’s a subjective timing). But there is a huge improvement. The new build has other changes as well as a generally increased feeling of speed. I believe it runs VBA macros direct from Microsoft Office. But the only one I have so far noticed is a set of ugly gnome-like icons, a small price to pay for the removal of the horribly sluggish startup that was one of the largest and oldest warts on the program.

More evidence of Sun’s inefficiency

Friday, December 8th, 2006

John Naughton writes, apropos the latest version of Adobe Acrobat, which makes commenting easier, that “Those of us who work in the Open Source world know that one of the factors which makes companies wary of moving to Open Office is that they have built their corporate working procedures around the commenting tools in Microsoft Word.”

But not everyone who works in open source knows this. In particular, the Sun developers in Hamburg, who actually write about 92% of OpenOffice, don’t know, and don’t care, despite being repeatedly told. The annotations in OOo are dreadful. They don’t even word wrap. An issue asking for them to be improved was opened on June 27, 2002; I still get messages every month from people wating to be notified of progress. There is no progress. There has never been any progress. It is not even scheduled to be fixed. Fifteen Eighteen other bugs in the database have been marked as duplicates of this one in the four and a half years since this one was opened.

This seems to me a perfect example of the ways in which OpenOffice can combine all the disadvantages of open source software with those of closed source, commercial software. Because there is no financial penalty for ignoring what users want or need they are ignored. Because it is so immensely complicated to hack on or even compile, the users have no chance of fixing things themselves even if they are motivated and capable amateurs.

Browser truce

Friday, November 10th, 2006
I have been using Opera for nearly ten years now, and for most of that time it has been the quickest and best-thought out browser on any platform. It had tabbed browsing, keyboard control, bespoke ad blocking, and full indexed searching of email and bookmarks before anyone else. You can get all these things with addons in Firefox, but it is a tedious and disorganised process.

But it still has irritating quirks. The email client sorts and searches wonderfully, but it doesn’t do anything else. In particular, it doesn’t do MAPI and it has a pretty horrible editor. There are always sites that don’t work at all with Opera, and some, like Flickr, that don’t work well. So it is a bad habit of mine to run beta versions of Opera, in the hope that these problems will be fixed, and the latest weekly builds have had a couple of bad regressions. First came a bug in the regular release which wiped out the multimedia keys on the keyboard. Since the volume control on my amplifier is tucked way under the desk, this was a major pain. It was fixed after about six weeks, but the new build has a problem which stops it displaying compressed pages from the cache, which is even worse. Since almost all high-traffic websites routinely compress their pages, this means you can’t navigate back to them.

So I decided to give Firefox and Thunderbird a try. I can’t see any real advantages, except that they more or less work. Thunderbird has much better handling of addresses and so on than Opera, and, I suppose, better text editing, though what I would really like is a way to use a proper external editor. But I can’t get used to having to wait for search results to come up. I have been so spoiled by Opera’s instant, incremental search.

Firefox works with more sites than Opera, which is always an advantage, and one reason why I already used it. The Adblock extension is a little better than Opera’s version once you are used to it. It deals more efficiently with Flash ads.

On the other hand, the process of looking for extensions is confusing and when you get them they are badly documented; there is nothing which seems to work the way that Opera’s note-taking system does. The skins are mostly ugly and sometimes jump alarmingly about. I really miss one-finger keyboard navigation. Opera has a wonderful system for pasting text snippets into forms. Its search bar is easier to customise and in some ways better integrated.

I think Opera is probably doomed, on the PC at least, if it can’t get its AJAXy shit together. Everyone seems to be writing that for Firefox and possibly Safari. On mobile phones, of course, it is marvellous. Still, I will probably switch back when a version is released where both the back button and the volume control work properly.

But, really, we have reached the enviable stage where browsers are about as interesting as washing machines, and easier and quicker to replace when they go wrong.

It’s like Ubuntu: there’s no really compelling reason not to use it. Who would have thought that software could make so much progress?

The official OpenOffice blogger extension, on the other hand, is just horrible, at least with MT. Back to my home-rolled posting device.

Slo-ooo-w progress

Wednesday, September 13th, 2006

The rate of progress on OOo is glacial; every fortnight or so I get an update on issue 6193, about the utter uselessness of notes in Writer compared to MS Word, and that was opened in June 2002, since when nothing whatever has been done. But even glaciers move, when you’re not watching.

I downloaded, and played with the latest version of NeoOffice for the mac, and there is nothing bad to report. It just installed; just worked; just did everything. Even the python scripting worked, which was rather eerie. In fact, it should work better than on Windows, since it uses the system Python and not a separate copy.

As a final tribute to its integration with the operating system, the keyboard malfunctioned as it should under OS X (confusion among the modifier keys).

Neooffice is an unofficial project, which uses bits of Java to get round the horrors of X11. But looking quickly at the release notes for the latest official build, it looks as if they are about to add a tabbed document interface. I haven’t dared test this. On the other hand, there is a delicious new bug whereby, if you save a document twice in Word format, all the upper-case non-ascii European letters are replaced with squares. Yet it works when you only save it once.

I wish I spoke Finnish

Wednesday, October 26th, 2005

At least sometimes I do. I found from the referrer logs a page full of people arguing in Finnish about my rant against OOo. I doubt I am missing much. The arguments are almost certainly the same in any language, up to and including Klingon — I bet that “MS:llä on bugeja ja muillakin saa silloin olla? Erikoinen ajattelutapa vai onko Open Source liike niin MS:n perään, että bugitkin pitää kopioida :)” means there are lots of bugs in MS products, and they don’t get fixed either.

But there are in fact damn few bugs in Word, and they do tend to get fixed. The very least you can ask of a software upgrade is that it continues to do the things that the old version could manage. Then you can ask it to do more. Lots more, perhaps: there are about 5,500 requests for enhancement in the OOo bug database, as well as the 5,700 defects reported. But OOo 2.0 has at least three regressions (defects that were not there in 1.1) which make it harder for me to work.

  • outline numbering is completely screwed up — 26 reported defects in the last year.
  • moving through the text, or deleting it, a sentence at a time is broken, especially when you do it backwards. I don’t know how many times a day I get half way through a sentence and then delete back to the beginning, but it is certainly one of the most common editing operations in my life.
  • there is a bad screen flicker when using the outlining and navigation window.

Going by past experience, it will be at least six months before any of these are fixed. That’s not surprising, when there are roughly 1,000 bugs outstanding for every developer. But it’s not much a of a value proposition, even when the software costs nothing to download. Yes, the MS paperclip was irritating. But I can turn it off. There’s no switch to turn on the missing functionality in OOo.

With enough eyeballs, all bugs are invisible

Friday, October 21st, 2005

An interesting test case for open source boosterism is provided by the release of OpenOffice.org 2.0 this week. I have been using this product since long before it was usable, out of a mixture of perversity, stinginess, and vague anti-Microsoft sentiment. When I started, MS Word 97, which was what I had, simply could not print out a 60,000 word manuscript without crashing, and I still think OOo may be better for books.

I have written a bunch of macros for it which people find useful, including the word count for version one. I have done bits for QA, not only submitting my own bugs but testing other people’s. So I am in a position to judge at the complete uselessness of the open source “community” that has gathered around it.

I now know of simple, hugely irritating, unfixed bugs that are four years old. I’d give links to the examples, but the site is unusably slow this morning. But notes (or comments, as Word users call them) don’t have word wrap; and for reasons I can’t begin to understand, spaces typed at the end of a line won’t show.

Most depressing of all, I have watched new bugs being introduced, then going either unfixed or unnoticed. Take outline numbering, something which was badly designed in version one , with a hugely confusing interface that took me a couple of years to learn.

In version two, the interface is still the same, slow, clunky, and confusing. But there have been changes under the hood, so that it now doesn’t work. There are 28 bugs outstanding against outline numbering, the oldest first reported in June 2002; and work has started on only one of them, an ugly problem where new toolbars appear and displace the text. 26 of them have appeared since a change was made last autumn. Things actually got worse a couple of months ago, and, though I have made an effort to reproduce the three unconfirmed ones, the behaviour is at present so random that I get quite different failures.

I know why they’ve shipped it. It’s six months late. There are only about 100 people working on it, and they had, at last count, 5,721 bugs outstanding. They have got to ship something. But if any commercial company, let alone one in Redmond, were to ship a steaming pile of crap like this, they would be derided all round the world, and rightly so.

UPDATE: Finnish readers, see here

Posting from Ooobasic

Thursday, October 6th, 2005

Well, I will be if this has actually worked …

Which will enable all sorts of strange goodies, if it works.

The really important thing is that it is extremely easy to draw dialog boxes in OOo basic; whereas in Python you have to write out long descriptions by hand, but you can steal libraries from everywhere to do things like posting to blogs. So a python component, which is possible for mere mortals to write, can then be summoned from an OOo basic macro, which can also collect all kinds of information from its pretty dialog boxes, wizards, etc, and pass it through.

So far as I am concerned, this just means an easy way to publish things here, instead of printing them. But that’s worth having, for anyone who types as badly as I do.

OOo down the drain?

Tuesday, April 19th, 2005

In the Australian Computerworld article about there being not enough developers of OOo/Star Office, Ken Foskey is quoted as saying there are now 50 people working on the program for Sun in Germany, ten for Novell, and only four independent developers. So much for the magic of Open Source.

Erwin Tenhumberg replies that he won’t talk about numbers, but that the ratio is misleading. This caught my eye because three or four years ago, when I first started using the program, one of the German developers told me that there were over 100 people employed on it in Hamburg. I can’t run down the email right now, but it did appear one one of the public lists. We have been reading for years about Sun making fresh rounds of layoffs, without any specifics. Occasionally well-known and admired sun developers have vanished from these lists. But to go from around 110 developers to around 50 in four years must have had a bad effect on the project.

UPDATE: three or four Sun developers piled in on the lists to say that there were still plenty of them, and there had never been more than 50 developers in Hamburg at any time. The rest were QA people and other support trooops, if that’s the analogy.