Tuesday, 29 July 2014

The Next Generation Browsers.....That Havent Quite Happened

Way back when in the early days of the www open internet, i started getting ideas about a super-browser which would be a personal oracle for content on the internet (in the original greek sense). By 2009 these had crystalised into a more specific concept of a personalised web-bot which would work in the background from your pc and deliver you the holy grail of relevant, salient and up to date information in a standard GUI platform.

This was no stroke of genius, all search engines have been striving for relevancy and anti spam since their inception, and personalised feed has been available since the pre www news groups and into the era of RSS feed (did anyone really use rss for more than a week?).  The ground was trodden but the trail had been blazed by the errant and often unstable marching elephant that is the do it all web browser.

Next generation to me in 2009 meant a super-info-deck outside of the social media circus, to which i was yet to become such an addict as I am today. All your favourite web sites would be scoured for new content, rss news would be screened for relevancy, meta engine searches would be compiled and pages automatically visited and spam screened, and the bot would crawl inexorably looking for highly similar sites and content to your favourites or those of current interest. Essentially a drag and drop lazy gui, which then did most of the leg work in sifting through the shit to get to the news and the stuff you are really interested in, at the same time you worked or even when the pc was on screen saver. More over server side or cloud systems could be employed to deljiver really well plucked content for you and present it in a deck which you could define as you will.

Today we are no where in particular nearer in terms of PC or standard tablet platform browsers: they are generic and have to cover the bases of all html and javascript while also being a cornerstone of consumer commerce and on line banking via secure https. All this and Bill Gates and competitors can't charge a dime, euro or rubel for the consumer software.

There were however many seeds and some still exist, of a more personalisable web browser interface with automatic content supply. For example rss as stated, but then google update alerts on searches and the loknger standing meta crawler pages.

What interceded in the potential for wonderful browser was of course the explosion of mobile devices and the two system tablet war. The outcome has indeed been personalised browers -plural- with uncountable apps which utilise html5 programming as a browser for specific source content, such as facebook or twitter. More open and nearer my concepts are the  newspaper crawlers and the multi social media decks.

In effect then the mobile device has become the interface by means of apps and some with their incessant alerts. I feel we are steered by them and use much of our time looking at paint drying in order to get juicy content and the whole process becomes distracting recreation like patience on windows 3.1.1.

Facebook in particular has its up-sides with news channels and feed filtering but it is irrritating to control the feed now and frankly i don.t want an algorythm deciding what is important for me. I have systematically dropped all meme addicts (bar one very sexy blonde) but then i miss all their good stuff and they get used to me being no likey them and no parles avec moi.

Mozilla firefox and app community has come nearer than many to providing a true post 2010 next generation browser, but they falter now and what was a fast, slim line browser seems to be baggier than IE v5. Opera has some plus points too, but the great white hope of Google Chrome has to be taken with the backdrop of consumer profiling and potential abuse by the US government.

Tablet browsers scare me a little too, especially the standard android one (which is wierdly not yet deleted in favour of Chrome presumably because they are device optimised) it allows for pop outs, split tabs and automatic downloads as soon as you stray down tin pan alley on the electric lady net.

The stage is hardly set and i bait no breath here : yes there could be a super deck, a non social media info oracle. Yes there could be a nice https only browser which is locked from all other browsing and client side app's. However no, it is in no-ones interest to fight against the major browsers and search engines and of course with a robot crawling basis you will get flicked off many sites or entire server banks or served spam from ghost sites.

Apps on mobiles are all the rage and they need to score fast on a pretty narrow benefit proposition, a USP or a brand halo effect, in order to gain critical mass and make money. When we start to think a big app, a  major cross genre of content personalisable browser which uses bot technology and perhaps some server side or cloud computing, then we are some time away. Not because of technology, which is all there, but because of the crazieness of consumer economics.

Thursday, 3 July 2014

Shimano History Proving the Unreliability of the Internet

...in particular of Wikipedia, not because of the concept of Wikipedia is so very flawed, it is just a little inherently flawed by it being really first to post. In corporate terms this is likely to be a professional web PR firm these days, but a few years ago it was out to the nerds run by the webmaster and some pretty amateur entries were punted for some corporations.

Shimano, case in point: in terms of bicycle components that is: The article is factually incorrect due to the "stub" nature of historical information on the internet. In fact the tradition of shimano is probably only held in dusty vaults under the sushi cafe at their HQ, and in some anorak's bedroom cupboard in form of old catalogues, pictures, boxes, reciepts etc.

On Wiki, both Shimano 105 and Deore are given a far later date for their arrival on the market. This is probably because shimano was not a major supplier even by 1980 after nearly a decade of producing good quality kit. So the aural or folk tradition that has transferred to the internet is tainted by what people actually remember, and that is tainted by asking Mr Average cycle enthusiast who is probably in their 30s.

Here though is one guy , Sheldon Brown,  seems to have a fairly honest collection of actual facts, pitted with some holes like the "W" notch ring machining which was a predecessor to hyperglide, where one tooth was sunken a little which generally promoted smoother downwards "derailing" , whereas hyperglide improves both due to the thinner, wedging machining to the lateral profile as well as the verticle height

The internet, and Wikipedia in particular, suffer from a terrible recency effect in what they present as fact. Disclaimer aside, it is just irritating. It is caused then by

1) Lack of Digital legacy history in  ASCII / HTML / XML / JPEG for most everything pre 1989 or even 1995 you could say.

2) "Experts" being contemporary, not knowing the full history,  or from say the USA, where as a product / Science / you-name-it may have developed for many years outside the US.

3) The search engine effect : recency, search word relevancy and links are pretty high at getting you on the first few pages of results, there after you are down in spam alley and fishing for other, correct key words. A search engine which snow-balled special key words to suggest new key words would be ideal, but probably reach diminishing returns in both quantity and how far back in time it finds results anyway.

4) Amateurs being the only ones who have time. WIki entries for "minority reports" and real corporate histories (which are often fascinating, especially for American technology start up successes) are in the hands of amateurs who are inaccurate and uninformed, or sometimes avoid the risk of revealing insider knowledge.

For Shimano, my three irritations are

a) That Deore has always been a mountain bike group set:

b) SIS was missing from its actual inception in the early eighties, or even late seventies

c) Shimano 105 was a far earlier group set than they claim

Taking a:

a) Deore was a rather pricey touring group set, with a finish in semi polished aluminium, probably available from the late 1970s. It was essentially a copy of european manufacturer's touring group sets, but priced below Campagnolo while above Sachs Huret in terms of quality too. Apart from the long cage arm derailleur front and triple chain set, the other innovation was ergonomic thumb shifters (which may have been an option on campag' touring sets earlier, but I admint to not being an expert)

The thumb shifters allowed for bar top shifting and were located on the inboard ends of the touring bike's bar-tape (or foam as had become popular in the 1980s) on the ram horn / drop handle bars. There may have been a deore bar end shifter too, which was popular in cyclocross, a kind of tab shifter which was mounted into the end of the drop handle bars on each side. Anyway, the rear thumb shifter received indexing before STI style integrated shifters came in, and the rear shifter retained the feature from the "down tube" old fashioned road shifter position, in having a ring to switch from indexing to friction mode. This actually was the preferred option for serious mountain bikers in the late eighties right through to 1992 or even later, well after STI and in particular, Grip shift had established themselves. I think it was the latter which killed off the market for them, as much as shimano doing a dirty, as with those indicator numbers we all though were so girly, but actually we quite like!  The front shifter was friction only.

In those days MTBs by in large did not sprout suspension forks for amateur racing because the forks were too soft and not reliable enough , so the bikes and riders took a beating. Also cables were probably not as well pre-stretched and not PTFE coated as they are now, so in the course of a hard race you could find your indexing had gone, or your rear "mech" was so clogged you had to switch to friction and more over, use your whole hand to shift the thumb shifter.

Some riders mounted the shifter below the bar, which has obvious ergonomic advantages, and this is where probably SCOTT got their idea for their odd bat wing under-bar shifters, which worked pretty well but it was easy to get your knee onto them when ejecting from the bike involuntarily.   Perhaps too Shimano got their idea for their under bar shifter from that too, but in any case it is fairly obvious to do it firstly as they did with the two levers, rather clumsy and always falling to bits, and then over to the ideal solution.

Grip- Shift (the birth-right of SRAM) was then the nerdy alternative, and I seem to remember it also having a friction only mode in the models from the early nineties.

One issue with the early Deore I remember seeing in the bike Shop where I worked then, Dales of Dobies Loan, Glasgow, was that the original front gearing had followed the european convention for 52/50 as the big ring, followed by a 44/46 and then a "granny ring" between 32 and 36 teeth. I can see the wisdom in this now, because of the principles of leverage - a smaller ring on the front gives you more torque, and also momentum, in that the amount of chain you turn and throw off is larger on a mid ring for the same actual "inches" of effected gear. So you reserve lower gears and higher torque per llb pressure on the pedal for the alpine ascents or the belgian "bumps" , and you have your 46/48 for rapid progress on the meandering, roller coaster back roads onto a wide back set where your effective 3rd gear sprocket is a 21 ( 28 / 24 / 21 / xx xx xx being usual low gears, with the xx higher gears depending if you were on the old 5 speed or onto 6 by the mid 1980s)

So successful was Deore in being a cross over set into MTBs that both Sun Tour and Campagnolo came out with copies, suntour if I remember correctly going for a cheaper and a better version, kind of LX and XT on each side of the standard, while campag had a pretty dear set, with their clunky non Hyperglide indexing onto 7 speed. I think it was the first campag group set of modern times which was brushed rather than highly polished. It was either lighter or rather heavier than Deore, I cannot remember but I know it differed.

b) Now to SIS. Shimano had a system called Positron ( not the rear differential trade mark) for indexed shifting which was catastrophically launched in cheap OEM bikes for the USA; with plastic levers and so on. In fact an american kid I knew must have had this in 1981 with the horrid plastic shifters mounted up on the bar stem. He had "clicky gears".

Anyway, it is claimed in Wiki and by Herr Sheldon that SIS came to light in the 1984 Dura Ace group set, which did so much to establish Shimano as a major player by being light and innovative. In fact I had seen a shimano 600 shifter from the early eighties with the SIS logo on it, and been told that it had been a fiasco by an experienced mechanic and road racer who worked at said Dales Cycles. 600 had been in fact offered to lower level pro's ( on the continent there are thousands who never make it to the "grand tours" ) with the SIS. Whether they had not perfected it, or more likely that those riders got little time from mechanics if the team actually had one, is for debate, but the mechanic did dig out an article on the system which looked dog eared and circa 1980.

The pre 600 EX group set was a rather scruffy looking affair, with a gun metal finish with speckled logo back ground, and I remember both my mate's Elswich Stag tourer having a mid to long cage version (probably a 26 mid length) and then seeing the SIS logo on a short cage version which came in on a bike.

According to the mechanic, shimano did not take into account stretch in cables, and the resulting difficulty in getting a correct indexing as the cable tightens when you go to the lower gears with more teeth on the rear block. This was solved by better quality "pre-stretched" cables and a longer adjustment barrell on the rear derailleur with it's own quarter turn cam adjuster widget inboard to facilitate quick road side adjustments (it may have been better to incorporate a fine adjust barrell or mechanism of some sort,  to the lever as they later did with STI and maybe Deore Thumbshifters)

I would not say that index gear shifting was "driven" by the mountain bike boom, where as STI was clearly a related development in terms of supply and demand. The earliest STI shifters I saw were dura ace and 105 , and both sets which came into Dales had an apparent "hand made" appearance to them, with some uneven metal working and polishing, and a milky finish to the aluminium incidentally.

Back to Shimano 600 being a rather cheap looking affair in the early 80s prior to the lovely much missed polished chrome like finish prior to the bastardisation to mediocre Ultegra-ness. 600EX was a beautiful group set, especially with dual caliper brakes, and when it was properly indexed around 1987 it was a better group set than Campag' Chorus for club riders, or the then fancy Athena which was lovely but still clunky in comparison to even the 105 SIS gear shift.  Prior to 600 EX the 600 rear and front mechs suffered from failures in the bushings around the riveted little pivot axles, however these could be re-engineered by any machinist worth their salt.

105 though..-

c) One-oh-Five. Shimano 105 is widely quoted on the internet as beginning with the rather innovative, and super light 1986/87 model, when Shimano produced an SIS group set that was lighter than the last Dura Ace. It featured SIS six speed, biopace chain rings (which I dumped) and soon into its life it had double pivot calipers, before 600EX got them. Also this had SLR, springy brake hoods, which did help the set up a bit - i had them onto nice 600EX single pivot calipers though because the 105 were frankly ugly and luckily out of stock when they built my 531 C machine. 105 had some other innovations to save weight in the bottom bracket and head set, and the hubs were super light. The finish was dull, brushed aluminium, but the next dura ace followed suit, dropping the fine polished finishe for a super light set for serious cyclists who didn't care about bling.

However, 105 has a longer history than that. I bought an old stock 105 derailleur , non indexed, from about the same production period as the gun metal 600 and it was a superior looking bit of kit, maybe a tad heavier, but nice and apart from not being a high chrome finish, the item was pretty much in line with Campag. The date must have been 1984 import latest, so maybe 1982 model, and I bought it in 1986 / 87 on an upgrade impulse.

The packaging box was chrome yellow with shimano's then blue, white, green stripe probably in the wrong pan tones.

I also saw a set of shifters (for some reason I bought campag, just to own some shiny campag) which also were better qaulity look and feel than the old 600.

There ends my rant about innaccuracy's, but I only pose my own half memories and half-truths to you.  

I would be pleased to get any comments which support this kind of chronology and so on of Shimano, while any criticism should come with non internet generation fact!

Post script It seems campagnolo suffer the same fate : i remember the athena groupset was introduced in either 86 in the UK or maybe a year earlier on the continent. It was non indexed and the last launched then without it. The lovely dual pivot brake calipers of chorus were in the set iirc, and these were nice - many say as good in chorus as the vastly expensive delta brake. Also the set had at least one other innovation which was at that time unique to athena: a special derraileur to rival perhaps Shimano's w-cut, or hyperglide introduced about that time and certainly to compete against indexed gearing which became standard on Shimano group sets in 1986. The first Athena derailleur had a geometry and mechanism which pivoted the jockey wheel cage backwards as the cable was engaged for an up shift to a lower (more teeth) gear. This meant that there was less chain on fewer teeth to disengage and made for a very elegant shift. I was told this was not new, but rather a more relevant introduction to what would have been a 7 or maybe only six speed cassette/freewheel over the five speed from history. I saw the athena group set on a bike at Dales Cycles of dobies loan, glesga, and tried the shift on a lifted back wheel and it was very nice if a little clunky sounding for some reason. At that time a machining company in Northenglandshire somewhere offered to machine cut any cassette or set of rear sprockets to "hyperglide" and I heard a few serious cyclists talking of getting this done to campag or older dura ace sets. Despite being poorer at indexed shifting for many years than the japanese, if you exclude the Armstrong fiasco years, campag have dominated the tour de france and many other competitions for the last six decades. They survived by making things of beauty and also eventually catching up in the innovation for top end group sets.