Since my mom started getting old 35mm slides scanned, all kinds of curious reminescences are bubbling up.

At work, in 1982
This is me, back when my parents’ business was just a room in our house. You can spot an Apple II (or, better, Apple ][), the box of a Vic 20 and, in the background, an Olivetti M20. In the next picture you can see that I was using a Sinclair ZX81 (connected to its fantastic “spark printer”).

Entity Relationship and PagesPlus

Yes, I do understand that yesterday post was a little cryptic. ;-)

At evectors we are working on a new component of the PagesPlus architecture designed to manage entities and relationship.

It all started by observing the flexibility of PagesPlus, which allows us to manage flows of content using different layers of tagging and plain queries to our aggregator and render the results on pages or widgets trough a templating mechanism.

Since most of the sites we design these days are built not only around content but also other types of elements (users, groups, products, companies, etc.), we started trying to figure out a model as simple as the aggregator to manage other types of data.

So, we invented the Entity-Relationship model (only to discover that somebody had already invented it much earlier) and we started building an engine which could manage in the most neutral way entities, relationship between entities and tags and make them available to our WYSIWYG page layout tools.

With this component (which we call Erm) we can define an entity and tag it as “user” and then define a relationship and tag it as “friend of”. But we can also create entities tagged “brand”, “company”, “product”, “group”, ect. or  relationships tagged “fan of”, employee of”, “owner of”, “belongs to”, etc, allowing us to define a lot of different sentences. Different types of entity can be linked to additional attributes, hosted either in our own databases or on some external service reachable trough an API, making the whole environment very scalable.

While developing these new components, we are also building a couple of real sites for real clients using this new approach (this is helping us keeping everything real and provides some very serious deadlines).

Yesterday’s video was displaying the tool which allows to create a query to the Entity Relationship Management engine (i.e.: find 10 entities of type user which have a relationship with this brand, sort them alphabetically and display their names and avatars) and render the result on a page.

Multi-language Social Networks

FriendFeed-1.jpgFrom the logo I see today on FriendFeed, I guess they just introduced localized versions of their UI. For the Italian and the Turkish markets.

I don’t care much, usually I don’t like localized versions, but the move totally makes sense: FF is quite popular in Italy, and from what I hear social networks are very popular in Turkey.

But while most social networking tool support multi-language UIs, what they are missing is support for multi-language users.

It’s what happens to everybody who speaks English and a other languages, and has friends speaking exclusively only one of those languages.

I ended up trying to write most of my stuff in English, because my friends on-line are a pretty mixed group. But I do realize that sometime I publish stuff in Italian, and while most people can live with it, I know that somebody is slightly annoyed by this.

I don’t think that figuring out a language of an item is very difficult these days, but I’m still waiting to see a social network that does this well, allowing users to receive only items in languages they can understand, if they want to.

As far as I am concerned: I’m happy to see some French, or German or any other language, crossing my feeds from time to time.

Time to change ISP

Last Thursday one of our servers, hosted with an ISP in the Bay Area, suddenly disappeared from the Internet.

Turns out we were among the many victims of the cable cuts down in San Jose, as we found out hours later.
I’m not happy with how our ISP handled this incident. It’s not just for the outage, I’ve been in this business long enough to know that shit simply happens and sometimes there’s not much you can do about it.
I could also probably live with the fact that nobody on my team was able to reach tech support on the phone for hours.
I could live with the fact that they were so badly effected, that even their websites and email services were down.
I could live with the fact that they didn’t put an alert in their voicemail message, the only form of contact from the company, describing what was happening.
I could live with the total lack of feedback on their twitter account.
I could also live with the fact that we only got a message from them after the network was back on line, nine hours later.
What I cannot live with is the fact that the in the post mortem message I received after the incident, the company CEO was blaming “vandals/terrorists” for what happened.
Using the “t word” as an excuse is simply ridiculous. Time to find a new ISP.

BlogLiguria – Il Secolo XIX

Last week we opened a new site based on PagesPlus: BlogLiguria.

The site is run by the staff of Il Secolo XIX, one of the most important Italian local newspapers very pupular in Genoa and the Liguria region. They are using our aggregator with a slightly different twist, letting anybody subscribe to the service and republishing automatically all subscribers’ posts, mixing them with content generated by the editorial staff and content gathered from public sources.

Il Secolo XIX-1.jpgA new PagesPlus feature which was introduced with this project is geotagging support: the aggregator is now able to decode the most used standards for geographical data tagging of contents, consequently next to the traditional “river of news” and “tag cloud” navigation, you can also browse the aggregator using a map.

Il Secolo XIX-2.jpg