It has been three months since my 25th anniversary, so I’ve been with IBM for 25¼ years. I was pleased by the responses I got to that post, but have been surprised by how many companies view relocation as either a hard, or even a soft, requirement.
However, so far one company has stood out: WSO2. They are geographically disbursed, committed to open source, and focused on wire level protocols. What’s not to like?
I’d like to see increased focus on scripting languages and REST, but I’ve worked with Sanjiva before, and in the context of a small company, I’m confident that one person can make a big difference.
It would be nice as WSO2 would get a clue about REST ... which they clearly don’t have right now :-)
Stefan, I’d love to hear more about your thoughts. Tungsten supports dropping a POJO into the server, and automatically exposes that as GET/POST with XML. I wouldn’t claim to call this REST in the pure sense of Roy Fielding’s thesis, but I think it’s what the average user is looking for.
Congratulations, Sam! And congrats to Sanjiva and other WSO2ers too.
I hope this signals that WSO2 is getting serious about the Web, and also less serious about WS-*. IMO, the last thing a startup needs is to be split-brained.
Sam, I need some help interpreting your log. The timestamps look like the output from time.ctime() - how do I read elapsed time (processing time) from the log?
That said, about 30% of the feeds that NewsGator subscribers are subscribed to provide neither ETag nor LastModified headers.
I agree ETAG support would be cool. I also think it would be good to offer a real Resource based model in Axis2 and Tungsten. But I think that most users are just struggling to get some simple XML communications going.
We’ve got a long way to go, but I think we’ve made the right starting place.
Ugh, sorry. I hope you don’t blame me for assuming you were being as subtle with the new job announcement as you were in making your availability known.
Gordon: yes, essentially it is time.ctime (though the documentation states time.time), so it is time in seconds since some date in the far distant path. The actual log message is after the network fetch and feedparser, but before any additional Venus processing or writing to the cache, so if you simply subtract the times what you are measuring is a mixture of times between two separate feeds.
That being said, some information can be gleaned from this data, and the conclusion is obvious: the fastest parser is the one that is never called. The difference is visually apparent, even at human speeds.
P.S. Since you are commenting here, I’ll take the opportunity to mention that I find it quite ironic that a site nominally for “Agile” programmers is still producing Atom 0.3 feeds.
Paul: there is no reason to be defensive. WSO2 has some solid products.
Many years ago, I was working closely with a company in Boston. I became friends with somebody who was roughly my age and, like us, he and his wife were just starting a family. In casual conversation, he asked me if I used cloth or disposable diapers, and without thinking much about it, I said something like “we use disposable, like everybody else”. His response: “you must know a different ‘everybody’ than I do”. And, thinking about it, Boston and DC in the late 80s were worlds apart in terms of social awareness.
This is good to see and I’m quite happy for Sam. These days, keeping your options open and choosing to exercise those options when the time is right is vital. When stuff like this happens, I always take a moment to reflect on my own situation. While...