Starting Monday, I'll be working for a client in Karlsruhe for a couple of months, optimizing some Linux C++ server software. I don't expect to be very available for GNOME stuff.
If anyone knows of any short-term apartments or rooms in Karlsruhe, I'd appreciate the help.
I’m back from North Berwick and a few days in Karlsruhe, trying to catch up on emails and To-Do lists.
I took the tourist boat to the Bass Rock for the first time. Next time I’ll try to take a zoom lense. The island appears greyish white from the mainland but that resolves, as you approach, to thousands (120,000 or so) gannets and their nests. It’s their realm.
I'll be in Edinburgh and North Berwick until the 11th of August, taking care of some family stuff, and taking a break. I'm not taking my laptop, but I am taking Robert Love's book.
Recent comments by Federico and Brian Cameron made me wonder: Are we ready to demand full documentation of any new APIs going into the GNOME Platform?
We usually try not to create feature-based criteria like this, so that we can actually release the software on a time-based schedule, as long as we are fairly sure that the important stuff will get done. So we don't demand that applications are accessible or internationalized immediately, as long as they can be eventually. But we have some limits – for instance, Platform APIs must be stable, not just likely to be stable some day.
The GTK+ team are 99% there. They document almost every new function, signal, and property, mark which ones are private, and when they were added. It works well. So, would any GNOME Platform maintainers rebel if the release-team said “No, you can't add that function (or add that module) until you've also written some gtk-doc documentation for it.”?
This week our GUADEC network guys (Joerg Hoh and Hanno Wagner and friends) are doing the same for ApacheCon, again in Haus der Wirtschaft in Stuttgart, and again with the enlightened support of the Stuttgart and Baden Württemberg authorities. That place really needs to just install a network instead of having one built and torn down every few days, even if Joerg and Hanno are really good at that.
I need advice from someone who really knows SQL.
This SQL command is simple and works, and hopefully the intention is clear. I get the actual contact name by joining on the contact id:
SELECT deliveries.departure_contact_id, contacts.name_last FROM deliveries LEFT OUTER JOIN contacts ON (deliveries.departure_contact_id = contacts.contacts_id)
However, what happens if I have two contact_id columns, and I want to get the full names for both? This is ambiguous, and incorrect:
SELECT deliveries.departure_contact_id, contacts.name_last, deliveries.arrival_contact_id, contacts.name_last FROM deliveries LEFT OUTER JOIN contacts ON (deliveries.departure_contact_id = contacts.contacts_id) LEFT OUTER JOIN contacts ON (deliveries.arrival_contact_id = contacts.contact_id)
Update: Thanks for all the comments. This seems to be common knowledge. You can use AS with LEFT OUTER JOIN to
give the join an alias, then use that alias name instead of the table name (even in the ON clause of the JOIN), like so:
SELECT deliveries.departure_contact_id, departure_contacts.name_last, deliveries.arrival_contact_id, arrival_contacts.name_last FROM deliveres LEFT OUTER JOIN contacts AS departure_contacts ON (deliveries.departure_contact_id = departure_contacts.contacts_id) LEFT OUTER JOIN contacts AS arrival_contacts ON (deliveries.arrival_contact_id = arrival_contacts.contact_id)
That's a simple change in a generic function in Glom, and I can use the relationship names for the aliases.
As the GNOME Foundation has now approved some funds for the European GNOME Events Box, I'm trying again to find a smaller 3Ghz PC for approx 700 Euros. I'd like a "small format" HP PC, like the ones I saw all over LinuxTag, but at 975 Euros, that's not better than a Shuttle PC.
A week or so ago I was in Linz, Austria, to do a little report for GNOME Journal about the Pangea internet/multimedia centre for immigrants and refugees. I was really surprised at the performance of GNOME using very old PCs as thin clients with a regular new PC as the server. They seemed as fast as new PCs.
A couple of issues came up:
I was a bit confused when arriving in Linz this time. Last time I was there, it had a vaguely eastern-european-style makeshift train station, but I suddenly found myself inside a multi-level shiny metal and glass complex. The tram used to be across the street but I couldn't find the street anymore. I followed the signs, and eventually figured out that I was indeed in Linz.
Glom now has an Image field type, for pictures. It's simple and it works.
Last weekend we visited Landshut, near Munich, for their Landshuter Hochzeit celebration. It only happens once every four years, and it was sold out in January, but a friend got us some tickets.
It’s apparently meant to be a reenactment of a royal wedding, and the accompanying festivities, that happened 500 years ago. Thousands of the local residents dress up in fancy medieval clothes and play assigned parts, so the whole town gets taken back in time a bit.
There’s a free parade down the main street of the old town, a huge area with fenced-in groups of medieval folks, like a human zoo, and a display of medieval entertainment. It’s less kitsch and a lot more fun than Oktoberfest, so I highly recommend it to visitors to Munich in 2009.
