froglogic / All / Musing on a Summer Afternoon

Musing on a Summer Afternoon

The end of the day at the office is nearing. Sitting at my desk, going through the list of people who have requested a trial version of Squish today. Of course several from the usual suspects, i.e. large
international cooperations. But what never ceases to amaze me is the sheer endless number of yet unknown shops of all sizes and domains that exist around the world.

I notice a high-pitched whistle filling my ears but can’t make out the origin. The harddisk of my machine started to sing its final song? A technician from the lab below passionately grinding a dental prosthesis into shape?

With our crew decimated due to vacation time it’s a bit more quiet than usual. Clearly audible are Rainer and Roberto fighting their last table soccer match of the day. Rob is busy with a new internal test library. Andreas and Frerich are discussing their major rewrite of our test development GUI. Just checked: no modal dialog blocking the user’s flow, yet. Aaron would be proud of them.

Looking at how much code I wrote myself today: not much. Busy with customer care, administrative tasks and design discussions. But that’s fine as long it allows the others to concentrate on their projects. We have a new product in alpha, another about to start later this year and of course many more planned. SO MUCH ONE COULD DO. If you want to join us on this trip and know how to hack in either (or all of) C++, Java and/or Python let me know!

Note to myself: reserve many hours of KDE coding time for the weekend. Investigate redoing the the tokenizer of KJS. With the recent performance improvements of the execution engine the parsing phase starts showing up in profiles. Should be able to do better than my simplistic implementation from ’99.

Next of the senses that gets stimulated: the cooking school downstairs has started their daily evening course. One scent more tasty than the other coming in through the open window. Hmmmmm. A master of the profession must be at work again. Getting extremely hungry, going home. Bye.



I’ve been following the KJS improvements with interest; thanks a lot for working on it 🙂

However, since hearing about Squirelfish I wondered if taking the same approach (very shortly, a VM instead of a parse tree. the Safari team has a great blog post explaining it) with KJS wouldn’t give a better code_speedup/developer_work ratio.
I understand this would depart greatly from the existing KJS codebase, maybe it’s too big a step for the available time ? What are your thoughts about it ?

Lovely blog, Harri! Those are real musings. It’s nice to have a look in your head.

Leave a Reply

Your email address will not be published. Required fields are marked *

Copy link
Powered by Social Snap