Tech H8rs

Marketing: Bits and Bytes

Wed, 19 Aug 2009 12:42:57 -0400

When talking about anything technology or engineering-related, especially when multiple companies are involved, standards are an absolute requirement to ensure that software and hardware play nicely together. These standards are often filled with technical jargon and field-specific terminology, and rightfully so, as they need to be very specific in what information they're standardizing. A blueprint that does not indicate its scale may yield a product far off from what it is supposed to be. Software, computers and consumer devices are no different, however there is an extra element thrown into the scenario, the consumers. This is especially true with the explosive growth of the internet and other high-tech devices floating around in every purse and pocket.Blueprints and industry standards are generally only viewed and useful by the engineers themselves. However standards, especially as of late, are being distributed to the mass population to help sell products and services, (all this jargon makes some people glaze over and just assume the company must know what they're talking about). How many boxes for wifi products sitting in retail stores boast '802.11N standard, get x-times the speed and range!'? The part that is not mentioned very often is the fact that increased range only applies if you use compatible hardware by the same manufacturer, as the N standard still isn't a technical standard and is still changing.

This rant however is not for hardware manufacturers upselling their products with technology buzz words that confuse consumers; this rant is dedicated to internet service providers, specifically Verizon and Comcast. They get to use a very misinterpreted tidbit of information, the terminology for data, ie: bits and bytes. Many IT-inclined people are well aware of the definitions of bits and bytes, their differences, and their connection to bitrates and bandwidth. For those of you who do not know about this IT jargon however, here is a quick computer lesson.

* Bit - The smallest part of data. Is a tiny part of your hard drive that contains either a '0' or a '1', on or off, etc..
* Byte - 8-bits that are grouped together to form a letter (essentially).

Examples:
* A = 01000001
* B = 01000010
* C = 01000011

Each letter is created by combining 8 bits in a specific arrangement. So, when you read a word on the screen, the computer is actually using a long string of 0's and 1's in such order to represent that word. If the word is "Hello", the string in the computer is "01001000 01100101 01101100 01101100 01101111"; 40 bits of data to represent 5 letters.

* Mbps - Million bits per seccond, 2Mbps is 2 million bits of information transferred in a second.
* MB/s - Million bytes per seccond, 2MB/s is 2 million bytes of information transferred in a second.

* 1Mbps = 0.125MB/s
* 1MB/s = 8Mbps

With this in mind, network speeds are all listed in bits, the smaller of the two; file sizes are listed in bytes, the larger of the two. Therefore, a connection speed of 1Mbps does NOT mean you can download a 1MB file in a second, but 8 seconds instead. Same applies to the connection speeds that ie: Comcast provides; they offer 5Mbps and 12Mbps speeds in one area, I'll use the faster of the two in this illustration. 12Mbps means literally 12 million bits per 1 second. Blazingly fast, right?.. well... on the 8-to-one ratio, that would give 1.5 million bytes of actual data per 1 second. Plus there is a network limitation of about 80% utilization cap, as the other 20% is used for background communication and overhead just to send the information, so knock that down to 1.2 million bytes per second (MB/s).

Another limitation with cable is the shared access, so it's not that you get that 1.2MB/s speed, but a group of people get a chunk of the overall trunk'd connection. (A trunk is a term for combining multiple lines together.) So, if the cable trunk supports speeds 10x faster then their fastest connection, (your cable modem will hard-enforce the speed that you are "allowed" to get), but you have 20 people connected to that trunked line, then your speed ranges between 1.2MB/s and 0.6MB/s, all depending on how much your neighbors are using their internet. This limitation does not apply quite as much to fiber connections such as Verizon's FiOS, as you get a dedicated connection of up to 100Mbps to their data center, (that's when the trunking occurs, but it's generally much higher then cable's trunking speeds).

So, as a breakdown on the two different common internet service technologies. With a connection speed of 12Mbps and best possible speed, it would take someone 2.5 seconds to download a 3MB song, and 84 seconds to download a 100MB video; much longer then the presumed 'couple second download time' that most people expect.

This falls directly into the main element of this rant, (aside from Comcast's marketing BS), is Verizon's marketing BS. In a recent radio commercial by Verizon, they state that with a 20Mbps download and upload speed, you can "upload a 20 meg photo gallery in a second". For this reason alone, someone needs to bitch slap whoever approved that advertisement. Based on the method for calculating actual network speeds, (Z / 8 * 0.8, where Z is the bitrate), 20 / 8 = 2.5 * 0.8 = 2.24MB/s. 2.24MB/s transfer speed would yield just shy of 9 seconds to upload or download a 20Meg file, not even remotely close to their claim. If however they were talking about a 20Mbit file, (which would be absurd because no one uses bits to indicate filesize), then it would take 1.25 seconds to transfer, much closer to their claim. This would actually only be a 2.25MB file under that premise. [If anyone happens to have a recording of that commercial, please post it..]

A direct relation to this is going to a car dealer and buying a car under the premise that it can do 0-100 in 5 seconds and has a top speed of 200, only to realize after-the-fact that the asshole who told you that was talking about km/h, even though you both live in a country that uses MPH for speed.

SVN Bugzilla Connector

Mon, 22 Jun 2009 01:02:54 -0400

One huge drawback with Bugzilla, and many other bug tracking software, is that they lack an SVN connector. This would be a snippet of script that would make a new comment under a bug, or close/resolve the bug automatically, upon a developer including a small bit of code in the commit message. This is usually something such as 'Bug#123 or Fixed Bug #123, etc...

This led me to some frustration when Bugzilla did not have one released. They have a workaround where you can redirect an email address to a perl script, which will then kick off the comment, but then you have to take the trouble of installing/configuring mailman, setup the correct aliases, etc... All-in-all, it was far too much work to get a simple autocomment script working.

So, after doing some quick crash-course learning perl and looking at existing methods of the Bugzilla system, I have developed my own script to use as an svn connector. I have the script available at dev.powelltechs.com (my dev server), available for download. A readme is available here.

The instructions for getting this setup is relatively simple if you have any unix experience, (and chances are, if you have your own SVN repo, you're at least *nix competent). Just edit a few parameters in the script, then link/copy it as your /svn/repo/hooks/post-commit file. It should start working with your bugzilla install and svn repo right-away, no need for patching or configuring Bugzilla or svn.

Enjoy!

Windows Starter 7 :: Restrictions

Mon, 15 Jun 2009 12:22:00 -0400

Microsoft has recently announced updates to what will be restricted in their upcoming ~~update to Windows Vista~~ "new" version of Windows 7, specifically their starter edition.

They have removed the 3-application limit, which is good news for whatever unfortunate soul happens to have to suffer through the OS on a machine. However, they still have some rather unusual limitations in place.
Some of these restrictions I agree with, although for different reasons then why M$ is implementing them. The ones I agree, at least in part with, are mainly the media centre choices. M$ is excluding media centre from the starter edition, which prevents users from viewing DVDs, music, or any other media for that matter. This may sound like an absurd idea for an OS to do, but in the aspect of what an operating system SHOULD be... On OS should not be loaded with a huge amount of extra junk, nor should the OS dev team have to concern themselves with such software. Windows is an operating system, not a media suit. If you want quality media playback software, look into VLC Media Player; it's faster, supports FAR MORE codecs, more user-friendly, and loads lighter resource-wise. What M$ should do is sponser some development from them, seeing that they are OSS; then just bundle that optionally when the OS is installed. Saves them dev time, money, and the user will get a better product!

Another feature they are excluding is their Aero Glass, (no, windows7 isn't just Vista re-labeled....), which I agree with because it's really only use is providing semi-transparent window borders at the cost of tremendous GPU resources. This was also another attempt by M$ to copy already existing software (Compiz Fusion [formally known as Compiz and Beryl], and Quartz), but of course failed miserably due to a complete lack of features.

A few other features, which are arguable, include AD domain support, multi-user logins, and multi-monitor support. The AD support I see as they want corporate customers to purchase the more expensive versions, although this probably could have been done better via their EULA instead of hard restrictions. Multi-user logins probably isn't a huge deal, as many of the machines barely will have enough resources to support the oversized hog that windows has become, let alone multiple instances of it. Multi-monitor support however probably should have been included. This is one rare area where M$ did something right. Multiple monitors are usually picked up quickly and easily enabled, something just recent to the *nix variety of OSs, (at least if you're lucky enough to be running an nVidia card). I guess their thinking here is that if people can't play movies, (because of course there is no other media player other then THE Media Player...), they don't need to display anything on a TV either.

However, the last feature that I have to comment on is their choice for the desktop wallpaper. They have opted to prevent the user from changing it, same with sounds and color schemes. WTF is M$ thinking with this move, (other then the obvious 'irritate the user into giving us more money for BS glitchy software'). Despite the fact that the starter edition is designed to be the base limited version, this is still a bitch move.

Now, these arguments are all null and void in respect to the argument that multiple subversions of the same piece of software are redundant, pointless, and just end up annoying end users and aggravating techs who have to deal with the complete lack of administrative tools in their lower end releases. This however is another h8.

If you would like to read the full article straight from the M$ blog, Check it out on windowsteamblog.

Non-H8: Flex Builder on Linux

Thu, 11 Jun 2009 12:14:23 -0400

Companies are extremely cautious to embrace new technology, even if the technology is not so new. One such very common issue of this is the development of commercial applications for Linux.

Eclipse has long been an expandable development platform for multiple operating systems. As such, it's the perfect platform to develop a cross-platform commercial IDE on. Zend ported their IDE over to it using mainly PDT and other various OSS tools, and Adobe is utilizing it for their Flex language (the logic in flash player). Despite this, they do not fully support Linux as a full/stable release.

If you want to vote for the Linux version of flex builder to be produced in full, please take a second to head over to http://bugs.adobe.com/jira/browse/FB-19053 and vote for it.

Imagemagick unable to identify EPS files.

Fri, 5 Jun 2009 18:36:00 -0400

I experienced a problem recently where one of my applications which relies on imagemagick's identify script failed on certain files. When executing identify filename_blah.eps on Ubuntu Jaunty, it would segfault and crash with a rather length trace stack, starting with libc6.

======= Backtrace: =========
/lib/tls/i686/cmov/libc.so.6(__fortify_fail+0x48)[0xb7af9da8]
/lib/tls/i686/cmov/libc.so.6(__fortify_fail+0x0)[0xb7af9d60]
/usr/lib/ImageMagick-6.4.5/modules-Q16/coders/ps.so[0xb7f24fe4]
/usr/lib/ImageMagick-6.4.5/modules-Q16/coders/ps.so[0xb7f2154a]
/usr/lib/libMagickCore.so.1(ReadImage+0xce9)[0xb7d52209]
/usr/lib/libMagickCore.so.1(BlobToImage+0x157)[0xb7d21c57]
/usr/lib/ImageMagick-6.4.5/modules-Q16/coders/ept.so[0xb7f1a6b5]
/usr/lib/libMagickCore.so.1(ReadImage+0x1cb)[0xb7d516eb]
/usr/lib/libMagickCore.so.1(ReadImages+0x1b2)[0xb7d52a32]
/usr/lib/libMagickWand.so.1(IdentifyImageCommand+0x4cc)[0xb7c654cc]
identify[0x8048c42]
/lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe5)[0xb7a12775]
identify[0x8048a71]

This seems to be based solely in some combination of that version of imagemagick and the core libraries of choice. The version that fails to identify the files is 6.4.5.4.dfsg1-1ubuntu3, however when I switch to either the previous stable version of 6.3.7.9.dfsg1-2ubuntu3 or the newest stable of 6.5.3-2, the command executes successfully.

So, if anyone happens to get this same error on eps vector files, simply switching to a different version of imagemagick should resolve your issue.

Bug report about the issue on ubuntu's bug tracker.
https://bugs.launchpad.net/ubuntu/+source/imagemagick/+bug/359181

The source is easy enough to compile if you want to get the newest version, just a simple "./configure; make; sudo make install" suffices. If you prefer an up-to-date Jaunty-based DEB file, contact me and I'll send you one of my compiled ones.

'Standardization' of HTML elements

Wed, 13 May 2009 01:21:10 -0400

The term 'Standard' can be very open to interpretation at times. This is especially true when it comes to software and computer systems. Take M$ for example, their version of 2008 Office featured an "Open Standard" implementation of a format that surprisingly, no other company implemented or even accepted; they just happened to have "open" in the format name, so it was a standard therefore.

However, this rather brief article is about HTML and its standards; not particularly that OS and browser variations will yield different results with the same code, but this h8 is about something that every browser does similar, AND is part of the standard. The standard I've recently disliked is that of form elements and the usage of the 'readonly' tag. As per W3C's specification:

readonly [CI]
When set for a form control, this boolean attribute prohibits changes to the control.

The readonly attribute specifies whether the control may be modified by the user.

When set, the readonly attribute has the following effects on an element:

* Read-only elements receive focus but cannot be modified by the user.
* Read-only elements are included in tabbing navigation.
* Read-only elements may be successful.

This is extremely useful, but there's one limitation of it... it doesn't support select boxes! This is rather annoying when ie: you want to limit a select box to be changed only via script, but still get submitted with the form. As per the standard, there is no way of achieving this; and thus, even if adding the readonly tag to a select box, you still get no readonly.

[ warning, shameless plug ahead ]

The easiest way to circumvent this is to physically prevent the user from interacting with the component, and the usual method for physically preventing interaction... place something between the two objects. In my case, I ended up creating a plugin for jQuery that will render a semi-transparent div or iframe over the element desired to be readonly. This will prevent the user from clicking on the element. Tabbing to the element is handled by assigning a custom onFocus event, which triggers the blur event for that element immediately, effectively preventing user interaction that way. This method seems to work well on all the browsers that I have tried, including IE6/7, FF3, Safari, and Opera 9/10.

If you're interested, checkout a demo and source at my development server, dev.powelltechs.com/jquery.readonly and let me know what you think.

Cloud Computing

Tue, 12 May 2009 14:10:14 -0400

This trend toward centralized, virtualized, data centers is a bad idea. I'm not against Cloud Computing per se, but the incarnation championed by Google and Amazon et cetera really bothers me. For those who are new to this terminology, Cloud Computing is essentially like taking a bunch of machines and aggregating the power available via those machines into a resource pool. Then you can dole out resources to users as you see fit based any factors you care to use, i.e. subscriptions, user levels and what have you. Sounds great. Looks good on paper. But suppose someone at your cloud provider notices your data is offensive. Maybe you support the second amendment. Maybe you want to audit the federal reserve. Maybe the person at your cloud provider disagrees with you. They can flatline your access, or anybody else's access, to your data any time they want. "That can't happen. Google is an American company, with American values." I hear you, but look what they did in China. And our values are changing as well. People are ready to give up civil liberties for "security". Corporations will not protect us or our data from the prying eyes of big brother. Just like its a bad idea for states to let the federal government control too much, it is in our best interest to maintain as much control of our data as possible.

Microsoft imposing limits on max number of applications running simultaneously.

Wed, 22 Apr 2009 00:45:44 -0400

Hunting is a primitive necessary feature and survival technique for the human race. We have evolved complex brains that are capable of comprehension and learning to aid in our survival and proliferation of an entire planet. One useful bit of information that we as a whole have learned is when killing game for food or sport, it is rarely advantageous to impale yourself with the weapon instead of the game. This often leads to bad consequences and the losing of said game you were trying to acquire. This may be the occurrences of Microsoft as of recent with their operating system strategies.Most operating system developers strive for a better, faster, leaner application so you can do more with less, even though hardware is advancing quickly. Such is the case with Ubuntu; with Gutsy Gibbon (8.04), bootup was comparable if not a bit slower then the uninfected Windows XP installation. The friendly developers over at Canonical seen this as a flaw and improved efficiency in the next release of the operating system only a short 5 months (approx) later. This made the speed faster then an average clean WinXP install. Not satisfied being a little better, they increased bootup time even quicker in the Jaunty version that will be released this Thursday (April 23).

However Microsoft tends to think "out of the box", if you will. Instead of making the platform take less resources and perform better, they decided to have the operating system use nearly 5-times the HDD space and 5-times the memory usage to do the exact same thing the previous version did. This could be considered akin to a car manufacturer producing two versions of the same car, both have about the same carrying capacity, about the same top speed, similar overall look to them, but the new version has cruise control and power windows; these features along with slower acceleration and 1/3 the fuel efficiency. Consumers would boycot the new car and continue driving the old one, much like the battle between Windows XP and Windows Vista.

This mistake would be shooting your big toe off, you can still continue the hunt and maybe even get some game so you don't starve, but you'll be walking with a slight limp for the remainder of the trip. However, their new decisions may be the equivalent of hacking off your good leg during the trip.

Even the name, "Windows 7", conveys that Microsoft wanted to have a completely different kernel and operating system then what the previous version was, NT6.0 (Windows NT Kernel on Wikipedia). However, this is version 6.1, not quite the version 7 that they had originally praised. This action should yield them to have to name the new operating system Windows Vista Attempt 2, as that is what it is, Windows Vista with a few things disabled by default. In addition to their extremely shady naming convention, they are going to continue with the multiple-version convention that has so gloriously annoyed consumers in the past, featuring a new base version.

This "Windows 7 Starter Edition" will virtually be a gimped-down Windows Vista Basic with the UAC disabled and 3D desktop rendering disabled. Besides the fact that you'd have to pay an addition "upgrade" fee for this, you'll be able to do LESS with it. Microsoft has decided to enforce a gimicky 3-application limit in the operating system, preventing the victim user from opening Solitaire if the internet, IM, and email are open. Just think, Not being able to open a PDF you just downloaded because you also have email open! How would any consumer stand for having the operating system gimp what they can open I must ask. I personally would not have believed this, infact I did not when I first heard about it several months back; until I read Ed Bott's post on ZDNet.

And as a point of reference, at work, I generally have ten different applications running, one of which is Sun's Virtual Box running Windows XP and Windows 2000 for testing purposes; running 3-times the soft-enforced limit of applications on Win7 starter. At home, that amount is usually comparative or doubled, (and by the way, I run Ubuntu, which is efficient enough to be able to handle running 20 different applications simultaneously AND still have the copmiz cube!).

Internet Explorer memory leaks with jQuery

Mon, 20 Apr 2009 02:29:20 -0400

Web developers can understand subtle differences between browsers, but some are just so far out in left field that no one quite knows what the original developers of the engine were thinking when they wrote it. Such as the case many times with the infamous M$ Internet Explorer, aka Internet Exploder, aka Infected Explorer.

Recently, a coworker and I stumbled onto a bit of an issue with this regarding IE's memory usage and javascript. We were using the javascript library called jQuery in an online application, and so we need to keep it updated. After a recent update, we were getting a strange bug; of course only in IE.

Example of javascript code that created the issue.



$('#logoLockerSelection')

        .dialog({

            modal: true,

            autoOpen: false,

            width: 670,

            height: 400

        });

        

        $('.logoSelectButton')

            .css('cursor', 'pointer')

            .click(function(e){

                

                $('#logoLockerSelection').dialog('open');

                return false;

            });

The use of a jQuery Dialog window would be perfectly fine in Firefox and Opera, but when displaying the window in IE, memory usage would spike to nearly 6-times the original footprint of IE, from approx. 50MB to over 300MB with CPU taxed at 100%.

Screenshot of System Processes

This would continue for up to 40 seconds, then the window would open and usage would drop back down. I spent some time trying to debug the issue, my only resolution was disabling the modal option would prevent this lag. After discovering this, I disabled the modal option and all was fine again.

However, we needed the modal functionality on a different page, which I was not the particular maintainer for; that page was maintained by Anthony. After some time of investigating, I eventually found a post mentioning something about IE and javascript-controlled divs that have an opacity on them. "No way, CSS would be causing 250MB+ of memory to be used on a webpage," was my comment to myself as I walked outside for a bit of fresh air so I wouldn't destroy my monitor.

Sure enough, Anthony installed an updated theme using jQuery's Theme Roller, and the memory inflation issue disappeared. Moral of the story, (besides to immediately install Ubuntu/SuSE on any M$ machine you find), keep your CSS up-to-date along with javascript libraries! Else you'll end up requiring the memory used by World of Warcraft just for the user to open up color selections...

Epic H8

Wed, 18 Mar 2009 02:17:59 -0400

As fitting for any new site, an epic introduction is generally used; however, due to my inferior word-smith skills, (my assets lie in other areas), writings of such quality are rarely rendered. In such, I shall provide merely a simple introduction for myself and the idea in which inspired this site, for whomever may stumble onto this site, however you may find it...

[quote]"I don't hate you because of you, however my hate is an indirect derivative from the fact that..."[/quote]

I, a mere nobody in the overall perspective of everything, was formally trained in enterprise and corporate network administration and windows / AD server administration and have acquired rich experience in face-to-face customer service and hardware troubleshooting. After moving into a completely new geographic area and losing all my contacts and clientele, I could no longer rely on basic hardware repair to sustain any randomness and expenses encountered, so a real job I had to search for. This yielded help desk jobs, help desk jobs, and yet more... help desk jobs. Sitting behind a phone all day reading printed instructions from a manual is just a hair more thrilling then preparing various objects of consumption with a tacky uniform on.

Freelance was where I ended up pursuing, both in hardware and software freelance, both of which to the same client, by some strange twist of the Fates' string. The hardware aspect has since ended, as an opt for virtual dedi-centers yielded a lesser need of a dedicated administrator, but the software development I still do freelance, provided not quite as much.

Software, a rather curious development of technology; created to imitate physical objects, but relying on absolutely no basic foundational laws. Humans see an object sitting in front of them, on which a particular piece is a different color, shape, texture, etc... We are programmed to explore and experiment, so we push it or otherwise try to learn how to utilize it. Usually, components of a different composition do something or are required to stand out in some way. Same thing if the said object doesn't function, laws of gravity dictate that as long as we're standing in either a spinning structure or on a massive object, throwing this POS off a cliff will yield satisfying results, if even at most temporarily. Physical objects must abide by the laws of physics. Software resides in a realm without such basic structural laws. If the user sees a button that says 'Close', there is no guarantee that the button will in fact close its parent, or even respond at all. I shall end my start of h8 there, as this could be a good subject for another day.

However, software, the skill that I have taught myself out of boredom on a piss-poor slow connection on my cruisin' 66MHz Pentium years ago. I originally learned PHP because for some reason, it appealed to me in some sick and twisted fashion. It was something that I felt was a basically worthless skill, as who would actually pay for some simple lines of code that any monkey could do. Even up to my freelancing in the area which I currently reside, I did not think to pursue anything related to it. Despite this, when I was nearly completely broke and just about to pack up my car and head back home, (literally within several days), I received a call from a strange person at some head hunter place. She told me that there was a company looking for a skilled PHP dev, lo and behold as I found out, I can get paid for doing something I enjoy, (most of the time).

Out of the enjoyment and contentment of actually having money enough to afford rent and food, (albeit not much else), I have developed a particular skill... This I have developed not quite so much out of my native personality, but something else, possibly social boredom... alas, boredom/idle hands are the devil's play things.... too bad I'm not christian ;) From wherever the skill was acquired, I have acquired it, the ability to look at many common items, and pull out the negative sides of them. This can include why IE is the bitch of the browsers, why Windows products are arguably some of the buggiest, most insecure, resource whoring applications ever built, the downfalls of including external javascript tags in your webpages, the strange obsession Apple has with the letter 'i', little 'quirks' in their OS, why linux distros can't bloody agree on a single set of paths for file handling, Xorg inconsistencies, or any other wide-variety of subjects.

This skill is called simply... HATE, or h8, as it is spelled sometimes on this blog.

My coworkers and I have many times in the past had debates or rants about various technological downfalls and idiocies. We decided that it would be entertaining for us to make a log of some of these, what better version of a log for web developers to use then a web log. Thus the creation of this site. A place for us to publicly vent, rant, and h8 about technology that pisses us off, in addition to any other random oddity we may come across.

And, a closing quote, in programming, the asterisk key '*' is commonly used to refer to anything in a search...

"I don't hate you because of you, however my hate is an indirect derivative from the fact that you are in existence, which falls under the '*' category for my hate."