OpenSolaris

Well, it’s out. OpenSol­ar­is, that is. I thought it would be fun to indulge in a little fake his­tor­ic­al writ­ing (or dream­ing, some of it really out there)…

Even the people in the middle of it all, those who came up with the idea of mak­ing Sol­ar­is open source, those who came up with the idea of launch­ing a major ini­ti­at­ive by blog­ging, did­n’t real­ise what an impact this would have on the world. Even now, 5 years later, we’re still dis­cov­er­ing nuances, still dis­cuss­ing wheth­er this product or that sys­tem could even have been con­tem­plated without the boost provided by the OpenSol­ar­is initiative.

So what was so new about this? Two things spring imme­di­ately to mind. This was the first time a major oper­at­ing sys­tem had moved from closed-source to open source. A num­ber of organ­iz­a­tions had made such a leap in the years imme­di­ately pri­or to OpenSol­ar­is, not­ably the Moz­illa Found­a­tion (which cre­ated the Fire­fox browser) but those were all applic­a­tions. Linux (the main oth­er open source oper­at­ing sys­tem) was open source from the begin­ning. The oth­er was the launch-by-blog­ging. Sun Microsys­tems was one of the first major com­pan­ies to truly embrace blog­ging by any employ­ee as a means not only to get inter­est­ing inform­a­tion out there, but to build com­munity, and we can now see (and this was inten­ded at the time) the blog­ging launch of OpenSol­ar­is as being a nat­ur­al part of that com­munity-build­ing pro­cess. The 132 Sun engin­eers who wrote 215 000 words to launch OpenSol­ar­is was seen as a dis­ap­point­ment at the time by some, as a tri­umph by oth­ers. It was clear that this was only the begin­ning of what would become a wave of launch-by-blog­ging mar­ket­ing that was copied widely and is a stand­ard part of any product launch today.

What was not expec­ted was the sheer speed at which non-Sun developers took the oppor­tun­ity to cre­ate end-user applic­a­tions. Up until the launch, Sol­ar­is was con­sidered a cor­por­ate oper­at­ing sys­tem. Sol­id, secure, but noth­ing fancy that would tempt a con­sumer to switch. There were few applic­a­tions cre­ated by third party developers; sev­er­al hard-work­ing cor­por­ate applic­a­tions but little of interest to oth­ers. But by the end of 2005 there was an explo­sion of applic­a­tions, some duplic­at­ing sim­il­ar work on the Linux plat­forms such as new desktops and ports of vari­ous pop­u­lar applic­a­tions, and some that were tailored to the OpenSol­ar­is oper­at­ing sys­tem, tak­ing advant­age of fea­tures that were not avail­able else­where. OpenSol­ar­is became the slash­dot­ter oper­at­ing sys­tem of choice, and reports star­ted com­ing in of com­pan­ies being deluged with requests to port their applic­a­tions to OpenSolaris. 

Now, 5 years after that fate­ful day, we see OpenSol­ar­is being used as the base of gam­ing con­soles, bank sys­tems, even mobile phones! Oth­er oper­at­ing sys­tem developers and vendors were inspired to improve their offer­ings, and the sta­bil­ity and secur­ity of all oper­at­ing sys­tems has increased to the stage where teen­agers don’t under­stand adult jokes about blue screens.

Ok, now back to real­ity. I use Sol­ar­is on my laptop so some of the above is my dream — more choice in Sol­ar­is-based oper­at­ing sys­tems and desktops (the Sol­ar­is equi­val­ent of Ubuntu, for example) and in applic­a­tions. There are lots of applic­a­tions avail­able for Linux (Skype, for example) that could be por­ted to Sol­ar­is with (I assume) rel­at­ively small effort. The basic things do work — brows­ing, edit­ing office doc­u­ments, email. I can even use punchin (though not in con­junc­tion with my Debi­an fire­wall, since I need to set up ipsec tables and that requires updat­ing to sarge, then recom­pil­ing the ker­nel, then fig­ur­ing out the magic incant­a­tions. If any­one is using punchin on a Sol­ar­is box behind a Debi­an fire­wall and has hints on the best way to tackle this, please let me know!). I like using Sol­ar­is, it feels sol­id and stable and although it’s been a long time since I worked on a Unix-based sys­tem, so I have to refresh my memory on how to do a lot of things, it’s good to have the choice.

Comment Spam

Any­one else notice a flood (well, trickle) of com­ment spams? They are com­ments of the form “great site”, “glad I found this site” etc, i.e., all mind­less com­ments that could apply to any site, with senders that are poker or online gambling sites. Some­how they’re get­ting past Spam Karma 2 (which nor­mally catches all the spam), so I’m hav­ing to delete them by hand. It’s not as it I have a par­tic­u­larly pop­u­lar site either!

Pilates Reprogramming

One side-effect of work­ing in the soft­ware world is a tend­ency to use pro­gram­ming as an ana­logy for life. Just to prove this, when people ask why I do Pil­ates I use a couple of pro­gram­ming ana­lo­gies about how the whole pro­cess works. Pil­ates is a meth­od of exer­cise which I star­ted doing after tir­ing of the way my knee would slightly dis­lo­cate when I walked. I’ve been doing it for about 2 years now and although I’ve sprained my knee since, it does­n’t dis­lo­cate any more (yes, this is pro­gress!). Any decent Pil­ates stu­dio (I go to Dianne Miller Pil­ates) will tail­or not only the pro­gram but the way it’s taught to each indi­vidu­al’s needs. I’ve seen two cat­egor­ies of teach­ing, with dis­tinct sim­il­ar­it­ies to main­tain­ing software. 

First, there’s what you might call the fix­ing bugs mode (or maybe TQM if you’re more into acronym-filled BPR ana­lo­gies). Strength­en­ing the muscles around the knees in my case, and teach­ing my over-achiev­er delt­oids not to do the work that the rotat­or cuff and ser­ratus anteri­or muscles should be doing (lots of the Pil­ates philo­sophy revolves around mak­ing muscle groups do the work, not train­ing indi­vidu­al muscles).

Even­tu­ally you’re far enough along the path that the instruct­ors decide it’s time to change everything — some­what like deep refact­or­ing, or rewrit­ing the ker­nel. So right now I’m work­ing on chan­ging the way I walk, and I’m back to doing the really basic exer­cises at Pil­ates in a dif­fer­ent way. I’ve heard people who golf a lot talk­ing about rebuild­ing their swing which sounds like a sim­il­ar pro­cess, with sim­il­ar trade-offs to deep refact­or­ing. If you don’t do it, you don’t have any major per­form­ance gains. But rework­ing the way you walk, or the way you do exer­cises you’ve been doing for two years, or a pro­gram you’ve been fix­ing bugs in for five years, can be a big under­tak­ing. Per­son­ally I think refact­or­ing pro­grams is easi­er than repro­gram­ming muscle memory — soft­ware sel­dom spon­tan­eously leaps back to the old version!

Upgrading WP

Last week I upgraded the blog soft­ware to Word­Press 1.5.1 from 1.5. Here­with a few notes for oth­ers plan­ning to do this.

The upgrade was mostly unevent­ful, with only two prob­lems of any note, under­scor­ing that you really should deac­tiv­ate any plu­gins you have before start­ing the upgrade process:

  • BAStats does­n’t work, as it appears some of the func­tions it calls have been rearranged into dif­fer­ent files. The author seems to be think­ing of releas­ing a new ver­sion at some stage, so for the time being I’ve just turned it off. 
  • The feeds were empty, but search­ing on the WP sup­port for­um came up with a patch (update: no longer avail­able, or needed) which fixed the problem.

Of course, now the Word­Press team have released 1.5.1.1, which includes that bug-fix, so I just spent the last 10 minutes (lit­er­ally) installing that. The 10 minutes includes back­ing up the com­plete data­base, deac­tiv­at­ing all the plu­gins, installing, react­iv­at­ing the plu­gins, and check­ing the results.

All in all a rel­at­ively pain­less upgrade. I gath­er there are a lot of secur­ity fixes, which alone make the upgrade worth­while, and it also meant I could install Spam Karma 2, which is a big improve­ment over the pre­vi­ous ver­sion, both in being able to get rid of old com­ment spam and not have it clut­ter­ing up your data­base, and in its fil­ter­ing abil­it­ies. Both of these factors made upgrad­ing worth­while. Now I just have to find the time to write more!

Friday the 13th

I spent a cer­tain pro­por­tion of Fri­day morn­ing watch­ing the web­cast of the Scott and Steve show — the update of the col­lab­or­a­tion work that Microsoft and Sun star­ted a year ago. Lots of oth­er people have blogged about what was shown and the implic­a­tions (try Tim, Eve, Pat, Robin, Greg and Jonath­an for some dif­fer­ing Sun views; CNET, The Register, and Slash­dot for some out­side-Sun views).

I had a spe­cial interest in watch­ing the demo (starts at 17:19) as I did the CSS for the Sun part. What happened was, I’m new to all this iden­tity man­age­ment stuff, and asked wheth­er I could help in pre­par­ing the demo for the eGov for­um at the Liberty Alli­ance meet­ing in Dub­lin in April, in part to help me try to get the con­cepts straight. Pat said the inter­face to the demo could do with some work; I took a CSS that I’d done for my blog (really must move it into WP 1.5 format some time to reuse it!), spiffed it up a bit, and voila! a CSS cus­tom-built for iden­tity man­age­ment demos. So it made sense for me to con­tin­ue spiff­ing up the demos for this press event. Unfor­tu­nately you only see about 2 seconds of the actu­al demo on the web­cast and it’s pretty blurry (the web­cast shows more of the demo presenters than the demo they were present­ing), but Pat assures me it looked great on the big screens in real life.

I took what I’d done for the eGov for­um, made some changes to make sure it would work on IE 6 on Win­dows XP as well as the JDS browser on Sol­ar­is (which, being basic­ally Moz­illa, does­n’t have the CSS “incon­sist­en­cies” that IE 6 has) and then pinched the col­ours from the new Sun brand­ing. Pat sug­ges­ted using some of the Sun images to add some piza­zz to the site, while Tim made the fake com­pany logos. 

Joint demo devel­op­ment takes a lot of coordin­a­tion. We dis­cussed small things like which logos we should show (we settled on Sun + Sol­ar­is / Microsoft + Win­dows) as well as big things like the pre­cise script that would be used, which defined how many links should be act­ive, and how many dif­fer­ent web pages each com­pany needed to pre­pare, which determ­ined how much cod­ing needed to be done. And a bunch of oth­er stuff, of course. We had daily phone calls with­in the Sun demo team, and daily phone calls with the Microsoft demo team, just to nail down all those little details.

In the end, we got there, the demo looked good and worked, and, I hope, made sense to people. Spiffy CSS or fancy images aren’t much use if people don’t get what’s being demon­strated. The write-ups I’ve seen indic­ate that they did.