View from bedroom:
View from office:
Wouldn’t it be a good idea if the Debian project leveraged their Internet voting system to research companies?
Debian developers are probably the only community on the Internet that can identify themselves with GPG public/private key signatures.
Debian Developers after all are generally highly informed technologists, of opinion of which I at least value greatly.
Another way of surveying or proving identity on the Internet is blogs I guess. What would be good if I could pose a questions like:
Have you heard of the University of Helsinki?
Now that question above would be a link whereby other blog users could ping it back with their answer, like Yes/No, and perhaps a short commentary. If I could specify the range of answers, all the better.
Then clicking on the question permalink would produce a summary of the answers from bloggers received via ping back.
Of course I am assuming that a blog belongs to one blogger, and that person is sort of identifiable, which is workable IMHO.
I think this would certainly be an improvement over the Slashdot poll.
Another point that came out of the discussions is that, in case there was any doubt, Internet Explorer in Longhorn will not support XHTML or SVG. (Microsoft suggested they would need some significantly more comprehensive test suites before they started working on standards compliance again.)
I’m 26 today
Left to right: Giuseppe, Tapio, Marko, Teemu, Elisa, Timo, Lie, Kai(that’s me), Philipp and Paul
Why is Insert mode “i” and Undo command “u” bound so close together on the most common QWERTY keyboard?
The amount of times I have accidently undone my last changes while trying to insert some text is mind boggling. I guess I should enter insert mode with more style, like “a”.
Tsk, Esc, esc…
Use case is that I want my mobile to alert me when my poisonous mushrooms have boiled enough.
- Find mobile
- Unlock phone
- Scroll down menu… Ah Settings… I want to set a 5 minute countdown
- Hmm… is it Time and date settings…
- Gah! It’s nowhere here… lets go back to the beginning
- Menu, scroll down… Past Settings this time…
- Alarm clock. Yes! Exactly! I want to be alarmed after 5 minutes
- Alarm time…no… I want it to count down damn it. Where is the count down feature. No I do not want to set my fucking tone.
- BACK AGAIN to the menu
- Hmmm, could it be in Calendar? ... nooo
- Applications. Well, yeah, count down is an application.
- Select application, App. downloads, Memory… ok I want to select the application
- Opening… (empty)... WTF. HELP ME LORD. I JUST WANT TO SET A 5 MINUTE COUNTDOWN TIMER
- BACK AGAIN
- Extras, Services… Hmmm which one… I’ll shoot for Extras as Nokia don’t seem to give a hoot about my time so why not extra?
- Sweet JESUS! Countdown timer. Ok, lets select this…
- 00:00, is that minutes and seconds? I’ll assume hours and minutes
- 00:05, OK… this must be it… please…. let this be it…
- Note for timer? WTF? JUST TIME DOWN YOU FOOL
- J (I am not going to type in “You’re mushrooms are done”), OK, Timer running. AT LONG LAST.
6 or so minutes later:
- Hey, five minutes must be up now… why doesn’t that Nokia 3100 ring ?
- OH FFS… It’s on silence. Oh shoot me now, life is not worth living if this stuff is our future…
- Mobiles have consolidated PDAs.
- Mobiles are consolidating cameras. But that can only go so far as better cameras need longer focal length.
- Mobile operators are going to consolidate ISPs?
What are mobiles again? Smartphones?
There are some computing trends that (or my) English is having difficulty with.
I would say a mobile computing device can be easily carried in a pocket and (characteristically) has a long battery life.
But of course that can change…
A laptop is portable. Oh no, portable is a synonym of mobile. Big screen, big keyboard, crappy mouse, 5hr MAX battery life (typically 1hr30mins on my thinkpad).
A desktop (or is it workstation?) is neither mobile or portable.
I am battling here. Comments welcome !
Another trend which I find interesting is the purchasing of computing devices. As an end user I so do not want to build a PC again. It’s fun, but a whole bunch of stuff can go wrong. I find that all prebuilt PCs are sort of crap and not very social(or quiet) as a laptop. My thinkpad is something I don’t care to open and I trust in working, as I believe many of the components are over-engineered to survive typical rugged everyday handling.
Personally I’ll like to have a laptop. It’ll be mostly sitting at home, and I would take it with me when I travel out of the country, or attend geek meets. At work a workstation. Screen should be level with my eyes to support a good posture.
Mobile phone. I would like it to do everything. I think taking pictures with it would be a great feature for me, as I hate lugging around my Kodak. It would have to be fairly decent, unlike the piss poor cameras inbuilt into mobiles currently. Priority for that device is to move SMSes over to Email I’d say. And to prove the interface to make it “Nokia 3210” fast and usable, unlike the newer phones like my Nokia 3100.
Media. Tricky word.
Paper media, CD media, new media, Internet media, Web media. And then again… medium.
The paper medium, the CD medium, the Internet medium, the web medium. Hmm is this making sense?
a document will typically need a larger font on a computer screen than on paper
Is this true? Why can’t font size be relative? I.E. Scale!
Doesn’t this not mean that an author has to set a font-size for different devices in order to make sure the text is readable? That makes media types device dependent, no?
Does this mean I as an author should avoid media types? The onus should be on the device’s browser user agent to make the fonts a “larger font”, surely?
Isn’t the web a medium all by itself? Why does it need style sheets for representations in other “mediums” (or it “mediums”)? Can’t we just treat the web as what it is!? A new medium, instead of getting it confused with existing ones?
Every wonder why the text width of email is set to 72 characters? I have. But I did not think of it in the context of the web… until now.
Mika Raento started the ball rolling by suggesting I should be setting a maximum width to my text like his symbian page, so a line doesn’t go too wide and becomes unreadable.
When I resized my blog or say Ian’s for a bigger screen (i.e. zoomed out), I imagined the text was scaling for the larger screen size. The fact that lines were getting longer, meant that the text was taking full advantage (i.e. scaling) for the larger screen size.
The register posts also get longer lines when zooming out.
Although the lines are scaling UP to the larger screensize, it is not helping accessibility.
Why? Because this line length style guide tells that “normal reading distances the eye’s span of acute focus is only about three inches wide”. Wide lines (roughly over 60 characters) get difficult to read, less usable and hence less accessible.
Mika again pointed out to me that optimal line width depends on:
- reading distance
- font size
Ok, what does this mean for mobiles? I think this is good news. If we read within three inches, and the mobile screen is typically one inch square (128×128), then there is no need to panic.
For desktops with bigger screens (17+ inches now?) it has of course had an impact. That’s why banner sizes are 468 pixels wide and many sites operate complex tables emulating magazine columns. This now affects mobiles. How do you scale a two column (at least 6 inches) desktop designed webpage to a screen sized the third of a three inch column of a mobile?
So how can we solve this problem?
- Re-adaption (guess and pray).
- Let the author worry about it. (let him do a bunch of different style sheets for media types or just one style sheet for one device and forget the rest)
- Let the device user agent worry about it
I am in favour of 3… better UAs to solve this problem. Let the UA agent in co-operation with device vendor take responsibility for making the page readable and usable. The more pressure you put on authors to scale their text properly for devices they’ve never seen is unreasonable. Authors should be just writing content. Just think how good it would have been if browsers (or some other application) could help you navigate instead of crappy author’s menu struggling to fill in the void with inconsistent results.
UAs can solve the font size problem. The vendor is in the best position to determine whether a “120% is not as legible as the same font at 112% on a particular device.” One down. As for reading distance, I think it’s probably more of a text flow problem. It’s not an easy problem to solve but UAs are best positioned to do so.
I am reading an excellent post by Brendan on the non-world non-wide non-web. With WHAT WG launching, people crying W3C foul and TBL actually in Finland this week for his 1 million EURO prize… it’s all getting quite interesting to say the least !
While reading the comments of Brendan’s post I read a comment by Robin Berjon which struck a chord:
To me the "small, incremental improvements to the ugly web" approach will make FireFox/Safari/etc be to IE6 what XHTML 1.x is to HTML. Neat increment, mostly a flop. Please don't take this as an attack (heck, I use your stuff day-in, day-out) but I would really like to see where you see a path to success in that strategy, because hard as I honestly try I don't.
When better things come along, evolved products and people don’t switch to it. Then evolution fails.
The average person does not want to change unless it “is an order of magnitude better (in whichever direction).”
In the natural world things get killed off don’t they?
Well things are never simple. But if you like me have been at this web thing for awhile, can you confidently say all your pages are valid XHTML? Damn, I am sure as hell can’t.
Popups break the browser paradigm
Dialogs such as accepting a certificate or allowing some data being sent must be done in browser forms.
Lets take a look at how two leading browsers handle a 404 and a domain which doesn’t exist.
It does the errors within the browser. No popups. Great.
BUT it could be improved. They are inconsistent. And the 404 page provided by the browser could included a feedback form so a user could easily give feedback to the site maintainer.
Wrong, wrong, wrong! What an unhelpful default 404. That screenshot shows the confusion. Is it a 404 now, or does the domain not exist?
On my Nokia’s user agent, the error ”STOP page can not be loaded” just flashes for a second and disapears. Hardly enough time to take it in. Can’t go back to it as it’s a popup. And after comprehending the error, the error does not have any useful information whatsoever to debug the problem.
Today I have been at the Technology Awards over in Dipoli, Otaniemi.
I was the only person there probably under 30 and definitely the only one wearing shorts and a T-shirt instead of a black suit.
There was a couple of good speeches, one by Chris Llewellyn-Smith about energy topics and his hopes for fission. He has quite a few credentials, and if he is saying resources such as oil will run out I believe him. I think his figure was roughly 40 years from now(!).
Today I voted green for the EU vote as they are dead against software patents as I am. It pains me that those idiots are against nuclear power. As for the short term, it is the ecologically best way to go about meeting energy requirements.
Another good speech was by David Cox about DNA and related topics. One thing I was reminded about was that it isn’t just our DNA. It’s the enviroment we live in. So if I was cloned… Also he compares genealogy to astrology. Another interesting comment was how <50% of drugs won’t work the same for everybody. Oh well…
As for the rest… the Finnish Prime minister was really boring. And Sari Baldauf gave one of those “business presentations at a technology conference”. Why O why… zzzzz
Oh yeah, I briefly met TBL. He was going on about the RDF to a journalist from the Economist I think. I find it really hard to get excited about the semantic web nowadays… Hopefully I’ll get chat to him again.
I was a bit pissed off that there was not any good food. And that wireless internet was not working either. I was invited over to Espoo city council for dinner tonight, but no one could tell me if they were going to show the football match between England and France for certain. Hence, I am not going. There is no way I miss a game like that.
Summary: Millennium Technology Conference = old people with money
Dinner tomorrow at Helsinki city hall sounds good, but I need a “white tie”. Oh no…
There were juicy quotes today:
- “Should we spend more money on R&D?”
- ” It’s incredibly difficult to predict, especially for the future.”
I hope tomorrow will be better. Although I need to cram for a statistics exam on Tuesday. Argh.
England lost. I really didn’t get much sleep last night. I am stressing over my exam tomorrow.
In the morning I did go to the Millennium Technology Conference (Day 2), and watch a couple of panels.
First one was on data communications and one thing that stuck out most of the 4 speeches was this fella Keen who said were moving from a Text to a vision to a voice interface.
I say the voice interface is pooh. Maybe this is my backlash for stupid voice answering phones.
With text you can:
* Use it even in noisy environments
* Read it fast, read it slow
* Skip over portions, skip back
* Index and search
* Manipulate in countless ways
* Capture, by means of easily recording and therefore transcribe
* It’s more accessible to deaf and blind people
* Pipe, redirect in Unix
* Convert to any medium
* And so forth (there must be more)
Second panel was quite interesting. About making molecular production systems. Drexler really stood out. An economist gave a talk summing up why there has been so much progress in the 20th century. It’s because with Intellectual Property and patents. It’s because people with ideas have had the means to meet people with money, but with no ideas. Hmmm… I should have dared to ask him if he thinks the system is now horribly abused (software).
After the 3 hours I was absolutely shattered. I feel disappointed with myself for being so out of it, and for not asking any questions. But it’s almost rude to ask questions with all those old Finns about. Jeepers.
So I have come home. Had some sleep, and will now study survey statistics. I have missed a panel with TBL and Tim Sebastian, which I a bit gutted about. And tonight there is a “white tie” dinner at the City Hall. I don’t a have a white tie.
Yesterday I attended the Web inventor awards in Finlandia, after an exam. Once again a formal stiff affair. I relieved myself of the tension (music/food was WEIRD) by helping myself to a few champagne glasses. When it comes to my turn I hope people won’t be wearing ties and perhaps some young women to look at.
Today there is a seminar in Dipoli (Espoo) but I didn’t go (3.40 EUR one way and a lot of time). I feel uncomfortable asking questions with that audience. My questions aren’t the “What inspired you sir to create the web?” type. My style would be a little more agressive (and specific) and don’t want to ruin the moment for anybody. Besides my thoughts like “Is XHTML dead?” and “Why does RDF suck so much?” were well answered by the guys on #buncs.
I have a lot to think about and revise in my mobile web thesis now. You’ll have to read my thesis when it is finished and give me comments.
Since Nokia’s XHTML browser doesn’t give a toot about media types, and is able to over-ride previous style selectors, many authors could offer the millions of Nokia phones access (not to be confused with accessibility) to their content !
I assume your site typically has a two column float setup for post and menu like mine. Trouble is on the Nokia XHTML browser(Nokia3100/1.0 (04.01) Profile/MIDP-1.0 Configuration/CLDC-1.0), the two column layout makes the text unreadable as it is too narrow.
So you need to reset back to original readable values. This seems to work if it is the LAST style selector it will execute:
Have a look at this page’s source and default style sheet for a better idea. Menu should come after the post.
Notice the handheld is not for the Nokia phone. It’s for the normal (desktop PC) screen browser. Nokia ignores the media type, but the desktop PC’s UA should only listen to all and screen. Hopefully Nokia’s XHTML browser will support the handheld media type in future, or even better… able to render the very popular two column view correctly without budging it up!
The next problem is that page size must be small, hence I have limited the front page to the last blog. I need to sort out the menu to be better and lighter…
Here is my writeup of useR R conference.
Austria is a nice country to live in, and I really like Vienna. I planned to absolutely rave about it when I returned to Finland, but then as I walking home from a bus stop I couldn’t help but notice how quiet Helsinki is. And I live in the middle of Kallio!
Vienna has lots of amazing buildings:
Hehe… And it was super to see some old friends. Public transport was great. The useR conference !
vigentte(“grid”) when I get pissed off with normal R graphics placement.
I learnt interfacing with R is non-trivial problem. Security being my biggest concern.
I sat in a few Financial track lectures, and it was interesting to hear how 99% of investment traders base their trades on Excel quality statistics. All those city jobs for supporting M$, and that’s how they work… with blunt tools.
Insightful's Jeff Coombs giving a slick CEO talk was quite facinating. R users heard how S-plus makes products out of R packages. He was asking that R and S should become more compatible. And I think it was Dirk on behalf of the R community who asked... "What do we get out from this?". I wonder if Jeff knew what he was saying when he said we send bug reports. Whoopee.
Without a shadow of a doubt, the highlight of the conference was Brian Ripley’s presentation on Data Mining. It was great how he attacked the dubious synonyms of Data Mining, KDD and well… data dredging. He pointed how it came out of machine learning, where machine learning is statistics plus marketing minus assumptions. =) He also showed up how Biostaticians typically never have enough subjects. Finally he parted on some advice, that statistical linear methods are fine and don’t get tempted by the hip Neural Networks et al.
R is very interesting free software project. It brings together computer scientists, polictical scientists, financial/business, statisticians, biostaticians, seimologists, geoligists, weather forcasters and other random geeks. =)
However after testing an inline image, I have discovered the maximum screen size for my Nokia… is… come on… guess… this is proprietary development sweeties… don’t be lazy…. 127×126. The amount of time I have seen “File to large to be loaded”! What good are User agent profiles?
Services->Settings->Appearance settings->Show images->No
Currently I see device independent web engineering as follows:
- The typical device dependent way. WRONG.
- The W3C device independence WG way. WRONG.
- Style and content are seperated with an emphasis on the UA to do its job. RIGHT.
The UA on a mobile would do this job really well if users can easily:
* Disable style sheets
* Disable images
* Scroll better (WTF should a server pagenate?)
* Allow larger sized documents than they do now
Read my thesis... there is a long list coming
The UA would also need to somehow recognize columns layouts and not budge up text. Text must flow full width on a typically small display of a mobile. Columns do not work. Possibily some screen scraping needs to be done with legacy HTML.
I’ve used the GPRS service extensively for development especially considering I was told I have unlimited bandwidth for 5EUR a month, but that seemed pretty reasonable considering the really low transfer speed. I’ve happily used sites like BBC almost everyday with my phone, even when abroad in Sweden and Estonia.
Last week Friday DNA sent me a message in Finnish though they know I am English to tell about their MMS upgrade. WTF do I care about MMS. After several emails back and forth with their customer service department as their multi-lingual website has all the current information in Finnish, I managed that “upgrade”. Then my GPRS service stopped working, with a message “Subscribe to GPRS first”.
I rang up this morning to find out from DNA customer service that I have never been subscribed to GPRS services. Ah great, that explains why I couldn’t find it on my bill.
But, I want my GPRS service back! I love browsing the web on my mobile! I am doing development too!
Then I look at the price list:
- dna GPRS 10 – 2,90EUR activation, 4.90EUR a month for 10 (10!) megabytes. 1.50 for every megabyte over.
- dna GPRS 100 – 2,90EUR activation, 16,65EUR a month for 100 megabytes and 1,22EUR for ever megabyte over that…
Ok, lets look at a competitor, Cubio:
* GPRS, activation fee 5 € (0 € until the end of the 2003), monthly fee 1,75 €, GPRS-data 1,75 €/MB
No wonder no ordinary consumer uses these services, they are so FUCKING expensive. I can smash a 100mb limit in about 10 minutes on my puny 1mbit connection at home. It’s outrageous how they charge that amount of money per megabyte. What is worse, is that there is no way of telling at least on my phone how much I have transferred. Leaving any consumer a little wary of being burnt by an unexpected bill.
All these asshole mobile operators better wake up. Not only is their customer service legendary, but their product services are a complete joke. If I had the money I would join their game, and beat them at it. I really expect better here in Finland.
Take the comparitively new Cubio. I tried to subscribe to their services, but it failed. They didn’t answer an email I sent about that. So I stuck with DNA. Today I noticed their new advertising campaign in the metro looks like Apple‘s. How imaginative.
Some fonts may not be legible at 120%, while perfectly legible at 112%.
I have seen device dependent engineering where pixels sizes of each limited font size from Nokia’s Development Platform specification are found and coded into the style sheet. Font sizes in pixels is not device independent. Previously I was advocating relative sizes. Then again, I think it is better to specify an absolute size keyword like large or small.
If authors are silly enough to specify pixels, UAs IMO should convert pixels to relative sizes. (Probably done already?)
The next thing is zooming or text scaling as I like to call it. If someone can not read the normal sized text easily, then it is not legible. The UA in co-operation with the device should provide an easy to adjust mechanism like little wheel or switch on the side of the device to scale, zoom, resize the text so it is comfortable, legible and readable to the user.
If the device screen can only support 3 different text sizes (small, medium, large), then everything has to go large. Therefore logical/structural markup should not be done with the help of text sizes.
Btw I am not talking neccessarily about accessibility here. This isn’t about people with disabilities. For example I could be sitting comfortably in a train and willing to have a very small font to cram in as much information as possible. However if I was out and running about in town, I would want the font size huge to read easier.
Summary: If this text doesn’t look great, blame your User Agent!
How should you display logical text? Text which has say more meaning or importance than another piece?
- Bigger font size
Blink and marquee (in the mobile profile btw) are extremely frustrating for desktop users, but… they work great on a mobile. Mobile displays must be stupidly eye catching. I still probably wouldn’t use them. Time dependence = Accessibility violation
Font size variation does not work IMO. My Dad always wants the largest font on his palm. If an entire text is at its largest font setting then it will not vary, and hence lose the meaning.
Italics for some reason is a bit strange. Firstly is not implemented on most Nokia mobiles. Secondly, what does its typeface mean? Under CSS1 it’s just a style. I just looked through a newspaper. It is never used.
Bolding or weighting is probably the best way to go. It’s well supported and for me at least it is unambiguous. It is something important.
Logical markup colouring is tricky. Colour blindless is too difficult to cater for. But colour is highly effective and well implemented (a little slow to fill colour on mobiles mind). It is difficult to be consistent too with the scheme. Using possibly gradients (with the help of a background) for distinguishing headers whilst using a large font might be the way to go. Although gradients would have to be quite different.
Underlining! Argh. I hate how some mobile profiles implement it. Underlining on the web medium is for hyperlinks. Plain and simple. End of story.
Actually, I have renamed my thesis to “Web Engineering for heterogeneous mobile devices”. It’s a little more scientific sounding, although I have my fears about the word mobile. Perhaps it just be wireless small computing devices.
Variability of mobile phones is an issue, but all the same I do believe mobile phones are consolidating PDAs, "smart phones", cameras and even the humble wristwatch.
Before I start enjoying my freezing Juhannus summer holiday I must throw in my 2 cents about the hubbub surrounding web applications and mobiles. I would love to see it all work out but you have:
- OMA‘s XHTML MP for mobile phones and really really bad default Nokia user agents
- Macromedia’s (amongst others) Flash/Flex doing real plugin representing
- W3C in the background with flops like UA profiles…
- Microsoft’s Avalon grappling for market share
Anyway it’s a turbulent time and by no means a time of convergence. Another point I’ll attempt to make is that people are raving about the network as the computer, server is über lalala. If user agents continue to suck like they do and worse on mobiles, then we’re not going to move very fast. OMG, there are so many problems everywhere…. ARGH!! I need a holiday.
- Riesling – medium dry – Alsace
- Chardonnay – oak taste
- Blanc de blanc – dry
- Suave – Italian, dry
- Chablis – French, expensive
- Pinot grigio – Italian, dry
- Merlot – is not so dry
- Cabernet Sauvignon – full body
South African and Chilian if you like new Country wines.
I have about four workspaces. Typically one runs a browser. The rest are big black *uxterm*s.
I find fault with GNU screen. Firstly Screen does not have an active mailing list. WTF. That’s a bad sign.
Second thing I do not like is that Screen does not integrate very well with ion and itself.
If you run screen, hit the default bind of the window list(ctrl+a ”).
First by default the terms there are not easy to distuingish. They are all bash on my system. It needs user@host and pwd. Better still, mutt or top when it is running.
Additionally it would be nice if it showed windows of all the screens on my system. I am running typically at least two instances/sockets/processes of screen. I can’t switch from one window of GNU Screen to a window from another Screen socket. Interprocess communication would be nice. Even across hosts.
It also sucks that I can’t say get my mutt screen window and split or move it into an empty Ion window. Screen should somehow recognize the window manager. I guess Ion too needs to somehow know about windows in Screen too.
Perhaps I am missing a few settings. Please inform me. Right now I am a little dissapointed with my interface. Focus problems and don’t even get me started about copy and paste.
I hope the whole complicated mess of the windowing paradigm aka the desktop will be replaced by simple tabbed MDI of the browser user agent.
Final comment is that Gnu Screen and Ion have such AWFUL names. Screen gets confused with the general term of a screen. And ion is the substring of a lot of words. /me sighs…
Just implemented a couple of GPRS bandwidth wasters which have been tested on the Nokia series 40 1.0 XHTML browser:
Only one submit button and optionally one input field at the top is all you can do on these devices.
That means the interface can be abstracted quite easily.
New to mobiles? Take a look at the M-Services Phase II Evolution PDF. Lovely titbits.
Section 5.1 Digital Rights Management (DRM) Policy
Digital Rights Management is a MUST in future devices.
My tip for Nokia today is if you want to clamber back market share, make a phone which just works. One where I can channel content such as ringtones and images wherever I please. Otherwise you’re just going to piss me off, and when a device comes on the market which is DRM free I will snap it up.
Who is going to lose right now? All those businesses based on ringtones and little images. What a loss.
DRM discussion is happening with PCs, but it looks like they are quite successful in slipping it into mobiles.
I have a Debian SID box on a fast link in Helsinki.
I need to cover some costs so I would like to offer shells for 5 EUROs a month.
Please mail me for more information.