Opera's intranet says I started at Opera 2001-02-17, and my last working day is going to be July 12th 2013. That works out as 4529 days (end date included) – or 12 years, 4 months and 26 days. Nearly 12 and a half years! Wow…
And that's it –
Now, allow me to look back and forward a little..
Opera Software ASA is something that shouldn't exist. By all practical and financial forecasts what Jon and Geir started back in 1995 was an impossible venture – a tiny company headquartered in a very expensive place, competing with some of the world's biggest IT companies, selling a product the competitors gave away for free. Sheer insanity!
Jon used to keep a framed letter on one of the office walls in Waldemar Thranes gate – it was a letter from The Norwegian Industrial and Regional Development Fund, a government startup incubation fund, rejecting the application from the upstart Opera Software. The fund considered Opera's business plan too risky to support.. (Well, who could possibly disagree with them?) Nearby, another framed letter from a user in Germany, who had become so enthusiastic about his improved web experience that he had stuffed a few dollar notes into an envelope and addressed it to Opera Software, Oslo, to pay for a licence..
I guess the juxtaposition of those two letters was a source of both pride and philosophical guidance to Jon during Opera's early years.
Now, I wasn't really around when a small team of very dedicated and talented programmers made the first Opera versions happen. Jon himself, Geir working on CSS and layout, Yngve on SSL and security, a little later Lars Thomas Hansen and Jens Lindström on DOM and scripting, Karl Anders Øygard on layout – Opera was extremely lucky in that the quality of the early work attracted great developers and quality-minded people.
Personally, I discovered Opera back in 1997 or so, when I got internet access in the computer labs at University of Oslo. In those early days, neither security threats nor security measures were anything like today, so I could freely install software found online on any of the lab computers. Naturally, I started every browsing session by installing Opera on whatever machine I happened to sit in front of :). Soon I was so familiar with it that I replied to questions on user mailing lists, created an online FAQ page, owned several paid licenses.. and one day Opera's customer support manager E-mailed me and asked if I was interested in a job at Opera Software. About 12 and a half years ago..
And the web was a frontier. It was an implementor's rat race through incomplete standards efforts. Opera went through growth, chaos, freedom, stunning quality in the details we managed to pay attention to, yet bugs, bugs, bugs. Sometimes the only thing that had more bugs than Opera seemed to be the Web itself.
Early strengths turned into old and aged features. Opera had support for fixed positioning way before other browsers – then, one day the competition caught up, sites like Twitter and Facebook experimented with mixing fixed positioning and endless, super-heavy pages – and Opera's old and proud implementation naturally turned out to be woefully under-performing and scrolling painfully slow. Software feature life cycles..
Never mind the bugs; Opera's ideas, quality and potential still got people excited. It recruited globally and became an extremely diverse company. It brought in investors and got some serious money to spend on expanding the company (like hiring yours truly). It didn't pay employees well – certainly not by Norwegian standards, in our insane oil-income-and-property-debth-fuelled bubble world. The headcount still grew, with a low turnover as far as I can tell, by the strength of ideas, visions and the potential impact on the world. (Pro tip: one of Opera's secret strengths was hiring its fans..)
Not that I ever complained about the money. Heck, I was basically paid for something I had been doing for free anyway – user support, browser testing, analysing broken sites – and working part-time obviously means a voluntary pay cut anyway. I was still studying during those early years, still thinking finance like a student. I probably still think that way :). Well, I'm possibly the cheapest employee to send anywhere. I was hosted by family and friends on long-distance company travels trying to support important Japanese mobile projects. I've couchsurfed in Lyon for W3C TPAC standards meeting. I slept on the floor of the former Tokyo office when I missed the last tube, ignoring the suggestions that the company should host me at a nearby hotel. (Not to mention that other time when I walked from the Hatagaya office home to Futako-Tamagawa after missing the last train. Several hours of calm, peaceful Tokyo-at-night mood..)
Of course, management would increase the compensation levels every year or so. And while that was very welcome (having gotten myself a family to support and all that), it would also make me a bit wary: can Opera afford this – doesn't it add more pressure on us to "monetize" our users? Crazy idealist, I know.. On the other hand, I know former colleagues who left Opera even though they enjoyed working here, because the banks didn't give them a mortgage due to the low income. I guess that says something about Opera being frugal, something about the industry, and a lot about Norwegian costs of living..
People often complained that there was too much chaos at Opera – not enough structure and decision making. Within this apparent chaos, I've had a lot of freedom. I've worked from anywhere I wanted to be, anytime. My most expensive working hour ever was a session at an Internet cafe in Lyon, before I got ADSL at our flat there. I think they were really suspicious of people bringing their own laptop, and charged me something outrageous – 50ish Euros, I think. Certainly more than I was paid for that hour of work :faint:..
Ever so often there was also some restructuring. Some new manager or department. Some new focus area. Never mind. I was happy to be working "on the floor". If some "please, someone look at this" E-mail bounced its way from some board member through the CEO and layers of management: hey, this is the right inbox. The buck stops here :). No matter how structures changed, I was usually focused on the same work. The narrow focus and specialization over time helped me develop analytic skills – oh, all those small and big detective stories hunting through foreign code for the real, underlying issue..
This focus and partly also my "free range" style meant I didn't pay much attention to office politics or larger developments. That's something I actually regret. I regret not pushing harder for User JS-based extension mechanisms earlier. It would have been interesting to try to push Unite towards its real potential. I wish I had been more aware of political powers shifting and Opera changing before Jon left.
Because Opera has changed significantly. It has matured, perhaps. Maybe – just maybe – the interests of investors and shareholders are considered more important than they used to be? Now, I'm trying to not automatically imply that this is A Bad Thing. Investors have given Opera millions of their own money, enabled much of what we have achieved. Shareholders believe in us and bet on Opera's growth with just as significant amounts of money. It's important stuff – and I think the "new Opera" is going to deliver dividends and growth and such things. Heck, I own some shares myself and I have no immediate plans of selling them. And obviously, the technical magic at Opera was supported by some financial and legal magic which I was blisfully unaware of.
And yet.. I'm used to working in an idea-driven context. When Opera Unite was dropped, I made a comment to a colleague and thought he was joking when he replied "well, if you figure out a way to earn money on having Unite, I'm sure they will keep it". I was laughing. He couldn't really be serious: were we supposed to ship only features that earned us money from now on? Not develop anything for the excitement? How funny!
And then we suddenly dropped Presto/Carakan/Quick, and I admit this feels like a major and miserable failure. Now, I never worked on the Presto source code – I know some developers find Chromium code easier to work with. I myself had long been envious of Chrome's quality, standards support and above all the stunning QA and development processes that enabled Google's unstoppable and apparently regression-free stream of improvements. Having less site compat work to do recently even enabled me to contribute significant improvements to the XMLHttpRequest standard's test suite – as an example of what we might achieve when engine maintenance and support becomes less demanding. And the first time I fired up a Chromium-based Opera Mobile build and saw GMail and Google Docs load beautifully was bittersweet. Surrender and victory at once.
So yes, I think Chromium/Blink is a great engine. It might work out well for Opera (when Desktop people have had a little bit more time to re-implement more missing features). We suddenly have an open source engine – which is great – and possibly will get significant accessibility benefits too. While plenty of old users complain about the missing features in Opera 15, I have an insider's perspective and know that the Desktop team made great strides in a relatively short time – this suggests that the new core really is easier to work with and that the development may accellerate. Of course, I don't have a crystal ball, and only time will tell if we can get the new version to a point where it tempts old power users..
However, there are still lots of small features, polished optimisations, ideas invested in the old engine, which we're leaving behind. As far as I knew, we were still growing Presto user numbers, in real numbers if not percentages – particularly with Mini. The most recent Presto-based desktop releases had some scary bugs (related to OOP plugins, SPDY) but I though development nearly reached a stable plateau of reasonable quality. And the handling-hundreds-of-tabs efficiency of Presto is an aspect of quality I believe Blink is nowhere near achieving (and when insisting on a multi-process architecture may be unable to achieve?). So to me, personally, it still feels like we, on behalf of Presto, snatched defeat from the jaws of victory, and it feels like Opera isn't quite the ambitious, audacious, idea-driven company it used to be. It's a good time for me to leave. I hope the Desktop team – rigged for and intended to be the new Opera's innovation engine – will prove me badly wrong about the "less idea-driven" part. Meanwhile, I'll go and do some work for Mozilla, trying to improve site compatibility for Firefox on Android. I'll probably be more useful to the web there than at Opera, in the near future.
And maybe, just maybe Mozilla is an environment where ideas are still kings..?
So, tomorrow is my last working day, and this is the last post on this blog.
Thank you for reading, and for using Opera during all those years. Keep an eye on it going forward – I certainly will.