Teky Blog

This is a big enough problem that I like to repeat it from time to time … but this time, I’d like to use a road/traffic metaphor to explain why YOU SHOULD NOT SEND PICTURES OR VIDEOS VIA EMAIL.

In this road/traffic metaphor for explaining email:

1KB- 10KB = “cyclist” = short email
10KB- 500KB = “car” = regular email, possibly with a small document attached
500KB- 5000KB = “truck” = large email, possibly with one high-res picture attached
5000KB-29000KB = “truck hauling a full-sized house” = huge email, with multiple high-res pictures or a low-res video or slide-show

Road/traffic metaphor for explaining email issues:

• Regular emails range in size from less than 1KB (cyclists) to hundreds of KB (cars).
• Regular emails with a document or picture attached can be hundreds of KB (cars) to a few MB (trucks).
• Now we get to those emails with HUGE attachments: videos and collections of high-res pictures. They can range in size from 5MB to 29MB … several times larger than “trucks”, so let’s refer to them as “trucking a full-sized house down the road”. I think you can imagine how much traffic disruption that can cause.
• Note: The reason huge emails aren’t larger than 29MB is that they simply won’t fit down the “road” … currently email servers won’t let an email through that is 30MB or larger.
• Any given email needs to pass through multiple email servers, and their size maximums can vary.
• So we have many people driving full-sized houses through our streets, not knowing how large their vehicle is, clogging traffic, and not knowing if they’re even going to fit through the next tunnel. (!)
• On top of that, each vehicle (email) self-duplicates … so there’s a copy left in the sender’s storage yard, and there’s a copy that gets parked in each receiver’s storage yard … clogging up storage yards all over the world. (!!)
• And on top of that, there is more “junk mail” traffic than “legitimate mail” traffic on the roads. (!!!)

Does this illustrate my point?

So, PLEASE … be conscious of the size of emails you are sending, be considerate of all the congestion you are causing by sending large emails, and spend the time to learn how to send large files WITHOUT attaching them to emails. That will be the topic of my next blog.

I’ll admit, I just made up the term, but I think it is worth consideration. Are you the type of person who clicks many times when you don’t get an immediate response? (Let’s call this is a TPQ of 1.) Or will you wait hours before deciding the computer desktop isn’t coming back? (Let’s call this a TPQ of 10.) What are the benefits, or costs, of high, or low, TPQ? Can “being informed” avoid the problems of high or low TPQ?

First, what are the costs of a low TPQ? For the person, it can be anxiety and anger, which can result in anguish and high blood pressure. For the equipment, it can be increased demands right when it is working its hardest, but it can also result in damaged equipment from being overworked, or from the operator lashing out at it.

“What can the computer possibly be doing that is taking so long?” Well, it may be inefficient software running on old or overtaxed hardware … or the computer may be doing incredibly complex calculations based on data that is being retrieved from overseas servers on lines shared with your neighbours.

Being informed can really help those with a low TPQ. (But, believe me, it is still no panacea.)

It is good to know the relative speed of your computer (5 aspects), your local network (1 aspect), and your Internet connection (3 aspects) … so let’s look at these speed aspects.

Your computer speed can be broken down into its 5 resources: CPU (central processing unit), GPU (graphics processing unit), RAM (short-term storage), HDD (long term storage), LAN (local area network). All of these metrics are now viewable in Windows Task Manager under the Performance tab, and most of them can be monitored on other operating systems. Your LAN speed is typically measured by just one number in Mbps (≤100Mbps is OK, >100Mbps is better). Your Internet speed is often broken down into 3 numbers: ping time in milliseconds (10-20 ms is good), download speed in Mbps (10-1000 Mbps is good), upload speed, also in Mbps (0.5-15 Mbps is good). These LAN/Internet speeds can be checked using https://speedtest.net or similar. When things slow down, it’s good to monitor these aspects of your system that can affect the overall speed.

Having a high TPQ is generally a good thing, but your time is definitely worth something (if not money). Ignoring abnormally-long delays from your technology wastes your time, and doesn’t address the issues causing the delays.

Like anything, taking an informed and balanced approach is the best. Knowing the relative speed of your technology, and the nature of what you are asking of it, helps to identify what a reasonable response time is. If that response time is suddenly high, that can point to an issue that may have cropped up. Identified issues can usually be resolved.

Each of us now, has a long list of online accounts to keep track of, and for that reason, passwords have become the bane of our existence. Previously, I wrote on how to create unique passwords so you can potentially keep them in your head. That doesn’t work for everyone, so this article is about how to store them safely and accurately.

As a computer technician and a business owner, I get tripped up every day, not just with my own passwords and my company’s passwords, but also with many of my clients’ passwords. Nothing can ruin the efficiency of a day more than a forgotten password. That’s why I say this about passwords: WTF = Write Them Fastidiously.

You can write them in a notebook, or keep them in an app, but if they’re stored in a document that is accessible online, I recommend you either encrypt the document or, even better (and simpler), use a purpose-built app that encrypts them for you. Some great password managers that do this include Google, DropBox, RoboForm, LastPass, 1Password, and Keeper.

For each account that you have, you should record 4 things: Title, URL (web address), Username, Password. You can optionally record 5 more things: Entry date, Description, Recovery email address, Recovery SMS, and Recovery questions and answers. If you’re doing this on paper, write legibly and leave room for when you change your password, or simply cross out the entire line and write a new one. Don’t keep multiple entries for the same account!

In case the fields mentioned above aren’t clear, here is an example and some details:

And please, don’t just write them down … Write Them Fastidiously!

Computers need at least two kinds of memory storage: volatile storage and non-volatile (NV) storage. Both are measured in bytes … but more commonly in kilobytes (one thousand bytes), megabytes (one million bytes), or gigabytes (one billion bytes). Volatile memory is typically called RAM (Random Access Memory). It is volatile because its contents are lost when the power is turned off. Non-volatile memory storage keeps its contents, even when the power is off.

Before getting more technical, I’d like to share the metaphor I like to use for these two kinds of memory. Volatile memory (RAM) is like the top of a desk … a “working space” for the files you are currently working on. Alternatively, non-volatile memory is like a filing cabinet … a space for your files when you aren’t working on them (i.e. when the computer is off). During normal operation, files come out of the filing cabinet, onto the desktop to work on them, and back into the filing cabinet for long term storage.

Examples of NV memory are your system’s main RAM, your CPU’s cache RAM, and your GPU’s (graphic processor’s) onboard memory.

An example of non-volatile storage is a hard drive (HDD): a set of spinning disks that store data using magnetism. These have served us well since the 1950s, but in 1991 we started seeing a new kind of NV storage called a Solid State Drive (SSD). Solid state drives store data using non-volatile RAM … NAND flash storage made up of floating gate transistors. They are much faster at reading and writing data than spinning magnetic hard disks, and there are no moving parts to wear out or get bumped. They are very dependable and offer a very long lifetime by allowing you to rewrite them millions of times before they reach their end of life.

HDDs and SSDs can be used on the same computer; we typically store the frequently-used data on the SSD, and the rest on the HDD. Hybrid HDD/SSD drives (called SSHD) exist, but they didn’t quite leverage “the best of SSD with the best of HDD” like we had hoped, so they have fallen out of favour.

SSDs are more expensive than HDDs, per unit of storage … but well worth the upgrade to get a read/write speed increase of about ten times. If a computer has a fast-enough processor to justify the upgrade, migrating a computer to SSD is a very good option to speed things up.

I was recently asked to “fix something permanently”. The most polite response I could muster at the time was “I got a sliver again yesterday”. A bit snarky, I guess. Another response that came to mind was “there are permanent fixes, but you’re not going to like any of them”. Too snarky? I thought so.

Technology is complicated. Technology is ever-changing. The only way to avoid the issues altogether is to avoid technology altogether.

I can’t begin to describe the complexity of a single personal computer. There are literally billions of components and millions of lines of code. Subsystem upon subsystem upon subsystem. Knock a bit out somewhere (called a Single Event Upset, ref: The World is Hostile to Computers), and the whole thing can come tumbling down. I’m amazed that computers work at all.

So we do our best, selecting the most stable hardware and software, maintaining that hardware and software, and preparing for issues by having backups and backup plans.

The thing that I was asked to “fix permanently” was a Windows backup over a network to a network drive. (Note: Complexity can be measured in bits, where 1 bit equals 2 combinations, 2 bits equals 4 combinations, etc.) That situation involved a third-party backup software (22M bytes = 200M bits), running on Windows 10 (3500M bytes = 33B bits), running on an Intel processor (roughly 2B transistors = 2B bits), over a network (wired and wireless) with a dozen other devices, and connecting to a NAS (another 200M bits running on a 2B bit operating system, running on 1B transistors). With the combination of all those bits (because they all have to work together), the number of combinations is more than the number of potential slivers on Earth (5×10^45, or 5 followed by 45 zeros vs the number of potentials slivers on earth = 3×10^12 trees x the number of slivers in a tree = 3×10^12 = 9×10^24, or 9 followed by 24 zeros). The network backup worked for 3 years before a problem appeared.

[In the preceding paragraph, M = million and B = billion.)

Of course I could fix my sliver problem permanently, but I don’t like any of my options. I can live with getting slivers.

Gaming for Linux has come a long way in the last 10 years, thanks to the continuing efforts of the open source community. No longer just the realm of hardcore Linux tinkerers, with little, or even no work, most games will be able to run on a Linux PC, as long as it meets the hardware requirements.

Lutris (www.lutris.net) was first released in 2013, and is continually being updated by the community. It is an open-source game manager that uses a Windows compatibility layer called Wine in order to run Windows based games under Linux. While most games will run OK under Wine, using Lutris will allow each Wine session for each game, to be customized for that game. No matter who you purchased the game from, download or physical copy, Lutris will allow you to install and play, providing someone in the community has made a compatible installer. If one is not available, you can try the default configuration, and should still get decent performance.

In 2018, Valve, the creators of the Steam gaming platform, took the code for Wine, and modified it to become new software called Proton. Proton (www.protondb.com) became the Windows interface layer for games running on Linux via the Steam games manager. It offers a much smoother experience than default Wine, or Lutris, however it will only support games that have been purchased through Steam*. As of now, approximately 76% of their top 1000 games are supported through Proton.

In late 2021, Valve will be releasing the SteamDeck, a PC-based handheld games console to compete with the likes of the Nintendo Switch. Running on a custom AMD processor, with SteamOS (Valve’s customised Linux for gaming) their plan, is that anything available on the Steam platform, will be playable on the Deck. Should this prove to be half as popular as the Switch, then we can expect to see more and more games developers supporting Linux.

So far, neither GOG, Ubisoft, Epic, or Blizzard offer their game managers for Linux, so using them through Lutris is an unfortunate workaround until they do.

*Note: For those who are familiar with manually modding games, getting non-Steam games working through Steam/Proton rather than Wine/Lutris is about the same level of complexity. Doable, but not for those who are afraid to tinker.