Entries:
Comments:
Posts:

Loading User Information from Channel 9

Something went wrong getting user information from Channel 9

Latest Achievement:

Loading User Information from MSDN

Something went wrong getting user information from MSDN

Visual Studio Achievements

Latest Achievement:

Loading Visual Studio Achievements

Something went wrong getting the Visual Studio Achievements

Discussions

Blue Ink Blue Ink
  • The Internet, 2000's, IE6/7/8, Microsoft, etc. what exactly happened?

    , fanbaby wrote

    I used to think that the saga was caused on purpose by Microsoft to sabotage the web, after all a truly open web conflicts with Microsoft business plan. But then i remember "Never attribute to malice that which can be adequately explained by stupidity". 

    So what's your take on the "dark years of the net". Was it dark? Is it still dark? etc.

    ...

    Let me tell you a different story. In the late '90s, browsers were implementing tons of features, sometimes to satisfy actual needs, sometimes just to one-up their competitors. And so it happened that both the largest browser vendors at the time implemented CSS wrong. And made enough noise about CSS that people started to use it.

    Any reasonable standards body, under these conditions, would have recognized that their definition of the CSS box model had been replaced by a de facto standard and amended their definition. This would have made 99% of the installed browsers compliant overnight. But the W3C feared that that would have set a precedent that would have taken the web away from them, by people actually implementing browsers (oh, the horror). Stupidity took over, and the CSS standard stayed as it was.

    Fast forward a couple of years, the same thing happened. IE had won the war, it was at some ridiculous percentage, and Microsoft proposed VML. Again, the W3C decided they didn't care about vector graphics on the web, and decided to reject the solution already available on 99% of the installed browsers and go with SVG. Which is vastly equivalent, except that it forbade use of vector graphics for more than a decade. I don't know whether it was malice or stupidity; I bet it was both, with a pinch of the "stick it to the big guy" sentiment.

    That's my take of the "dark years of the net": it was a political match that we all collectively lost, in terms of features that could have been and weren't, standards made incompatilbe on purpose, long hours lost fixing bugs.

    Microsoft paid for (some of) its sins, the W3C didn't and it's unlikely it ever will. But the web will be a dark place until the W3C changes its ways and becomes what a standards body should be: a place where common sense trumps politics, good engineering practice trumps philosophy and the best interest of the end user (if not developers) is all that matters.

    Never attribute to malice alone that which can be adequately explained by stupidity and malice and shortsightedness.

    P.S.: FrontPage was "the web for the rest of us". It allowed you to put up a webpage without knowing a thing about HTML, CSS, JS. The result usually wasn't worth the effort, but that's what I think made the web truly open and accessible: the concept that the average Joe could actually buy a domain and put up his homepage with little more effort than it took to edit a regular document.

    That's good engineering: making stuff that's good enough, stuff that works and enables everybody else to do something that would normally be beyond their reach. That was BASIC, that was DOS, that was FrontPage. Those who complain that "it's not semantic" are elitist boneheads that may be fit for academia, but shouldn't ever be allowed near something that's actually useful.

  • WP7 and IMAP emails

    Yes, with the due exceptions (IIRC the CA role requires an Enterprise server, and makecert is better known to developers than IT). All of this, including OpenSSL, requires an IT dept that know their stuff, and when a company is pinching pennies enough not to buy a certificate that's certainly not a given.

    If you want a sense of what the state of the art is in the wild, just bing "renew exchange certificate". Just don't blame me if your hairline recedes by an inch.

    Again, I'm not saying you cannot make this work, I'm just saying that the defaults make things worse than they should be, for no obvious reason I can see. And make WP7 an unwelcome choice for many of my customers.

  • WP7 and IMAP emails

    , blowdart wrote

    *snip*

    Of course considering that SSL Certs are like $20 these days using a self signed one is a silly decision. Especially as that will also get used for S/SMTP so inbound emails will see it, and the sending system may decided not to send as it's not a valid trusted cert. (assuming everything is on the same host).

    I could never find anything that cheap, but that's beside the point. If companies are allowed to use a DIY certificate, it doesn't make sense to force them to trust yet another DIY certificate every year.

    And it makes even less sense to make you send a certificate to a device. Not exactly material for a "Smoked by Windows Phone" challenge.

  • WP7 and IMAP emails

    , blowdart wrote

    *snip*

    They don't have to expire every year you know. You just need to create them right Smiley

    I don't have troubles believing that. But since that's the default for Exchange Server, both for creation and renewal, that's the setup I invariably find. Hence the irritation.

    The excellent reasons why a company should stop trusting its own certificates every 12 months by default are way out of my depth...

  • WP7 and IMAP emails

    A word of warning: if your email server requires SSL, expect problems.

    I had my fair share thanks to self-signed certificates (Exchange Server and an IMAP server of unknown pedigree), and for other unexplained certificate problems (a Lotus Domino, with a certificate that appears to be valid and trusted all the way up).

    Needless to say, an iPhone will just ask you (once!) if you want to accept the certificate, install it and just work. With a WP7 you have to get hold of the actual certificate, email it to the device, install it from the attachment, reboot the phone.

    And since self-signed certificates expire every year, you never run out of fun...

  • These seem like compelling reasons not to work for Microsoft to me

    @1001001: Yeah, why not go back to - say - Renaissance and get Da Vinci fired for being an incompetent?

    Seriously, I don't think many of us would succeed against these guys (and gals), in 1975, using their weapons of choice. We had the benefit of decades of training on computing they couldn't possibly get, but we were also spoiled rotten in the process.

    The most advanced workstation back then was an 80x24 terminal, a crappy editor, and some 8-bit assembler. Manuals (the dead tree version) you had to contend with your coworkers, no internet to find information or ask questions, compile times that ran for hours. And all this to produce code that had to fit in a handful of KiB.

    We could relearn, but I doubt we would stomach it.

  • These seem like compelling reasons not to work for Microsoft to me

    @JohnAskew: I believe Maddus was referring to something different: there's nothing wrong in getting a job you aren't qualified for, as long as you are willing to bust your arse to learn and catch up.

    That's not the Peter Principle: that's what keeps the world interesting for the rest of us.

  • Interesting question

    @ZippyV: as I recall, the Prius I looked up on that site was rated about 66mpg (UK) or something like that... "Your mileage might vary" was never more appropriate.

  • Interesting question

    @ZippyV: 60 mpg is nothing to write home about these days. Leaving alone pure electric for the current range problem and battery lifetime, there are countless diesel and petrol options that easily exceed 70 mpg (there's a VW Golf 1.6 TDI that makes 74, for istance) and you can exceed 80 mpg on some models (source: www.nextgreencar.com ).

    I would have expected hybrids to do a lot better in comparison, and given that diesel engines are more efficient than their petrol counterparts, I would have expected that there would be a lot of diesel/electric hybrid cars by now.

    100mpg isn't unrealistic for one of those and that would be an excellent stopgap while we mull over how we get to recharge and manage batteries...

    EDIT: for what it's worth, diesel fuel is also extremely safe: its flashpoint is high enough that it's not even rated as flammable.

  • Thought provoking article by Rob Pike that C++, Java and C# (especially) devs should read

    Ok, as an old time C++ programmer I'm biased. And opinionated. Take with a grain of salt.

    First of all, the whole point about the bloat of C++ is a fallacy. True, C++11 did result in adding a lot of features to the already gargantuan specs, and it couldn't be otherwise given the unprecedented codebase that must not break. Noblesse oblige. Yet, once you start writing "pure" C++11 code, you end up using a relatively small subset of the specs, which results in a lean and clean language, without renouncing any of the power and efficiency of the language.

    Second, there's the preposterous idea that a GC is needed for parallelization or that it makes a language intrinsically better. There's nothing wrong in using a GC, of course, I just don't see it as an advantage, not against modern C++. Conversely, lack of deterministic destruction, RAII and the like is a serious limitation for a "systems programming language", where fine-grained resource management can be paramount.

    Dissecting each and every feature of the language is beyond the point, let's just say that the language doesn't innovate enough, with respect to C++, to generate more than a mild interest; on the other hand, it has a number of "features" that I just find irksome, when not altogether misguided. Postfix typing, overloading well-known symbols from other languages with totally unrelated meanings, implicit interface inference (interesting, but that's an accident waiting to happen), the egregious short-sightedness of using case to determine visibility (not to mention cultural chauvinism... ironic for someone who has developed UTF-8). Just to name a few.

    And all of this syntax sugar just to save keystrokes. Might have made sense in the '70s; nowadays it's the task of a decent IDE to take care of that, not the language design.

    Which takes me to the most important reason why I haven't taken up Go for a ride beyond the online demo: features (or lack thereof) are but a facet of what makes a language the right tool for the job. You must factor in IDE support, ease of debugging, unit testing, documentation, libraries, tools. For a language that is apparently all about "getting the job done", Go doesn't offer any of that, not up to modern standards anyway (gdb? seriously?).

    I'll leave it at that, for brevity <g>