Technology facts you thought you knew

[Paul]

• WHOIS is a protocol, not a series of websites. You don't necessarily need a website to use it. There are WHOIS clients, just like there are email cients and web browsers.
• The Web and the Internet are not the same thing. The former is a subset of the latter.
• FTP isn't the only way to upload files to a web server. There are more secure options, but, sadly, they're not as easy to use.
• Mbps and MB/s are different kinds of measurements and don't mean the same thing. It's the casing of the "b" differentiates them, not the slash ("/").
• Almost all passwords can be cracked quickly. If you're concerned about security, be sure to use a different password for each website and a good password manager.
• Don't use JPEG for images on your website. It's a lossy format: that means the image will look worse than if you use a lossless format, such as PNG. This is especially apparent on devices with high-DPI screens, such as smartphones.
• Photoshop isn't always the best tool for image editing. It creates raster images: pictures made of tiny dots. For computer-generated images, vector images are preferred; they are made of lines and shapes, so they don't lose their quality when they're enlargened.
• Large shared hosting providers often host thousands of websites on a single server. It's possible to see who else is on your server, and sometimes even to edit their files.
• Globally, Android's market share is much larger than iOS', and is growing. iOS' market share is shrinking. However, in the US, iOS is pulling ahead.
• Despite being a very successful company, Google has started plenty of big projects that ended in failure. It happens to the best of us.
• Mining Bitcoin is more or less the technical equivalent of cracking passwords--except the latter is (usually) much easier.
• Apache, while easy to use, isn't the most powerful web server. Try Nginx ("engine-x") or Lighttpd ("lighty") for important websites.
• The United States is producing fewer and fewer Computer Science graduates each year. This is a big problem, as there are already more jobs than qualified professionals.
• You think gas is expensive? Inkjet printer ink costs around $4,000 per gallon. Invest in a laser printer: you'll save a lot of money.

 

[RU]

Windows 8.1 rocks!

 

[Joseph Green]

Mbps and MB/s are not the same thing. The capitalization of the "B" matters. Mb/s and MBps are misleading and therefore shouldn't be used.

Additional details:

• Mbps = Megabits per second.
• MB/s = Megabytes per second.

There are 8 bits in a byte, so 50 MB/s = 400 Mbps.

To convert from Mbps to MB/s, divide by 8.
To convert from MB/s to Mbps, multiply by 8.

 

[enlytend]

5. FTP is extraordinarily insecure, and is not used by enterprise professionals. To make it a little better, most FTP servers let you connect with FTPS, which is secure. However, most FTP servers are constantly getting attacked by automated hacking bots, so it only solves some problems.
7. A password that most people would consider secure can typically be cracked in a matter of seconds or minutes, even with strong hashing algorithms. Passwords should be a randomly generated combination of letters, numbers, and symbols at least 17 or so characters long

5a and 7a:
Bots attack the most "likely" default settings the hardest. You can reduce your vulnerability a lot by NEVER TAKING DEFAULT SETTINGS where you set up a web site, email account or piece of software. Have cpanel hosting? DO NOT use the default user name or root directory name for your sites.
Using wordpress? Never EVER name your administrative user "admin", and preferably use other custom settings like unique table prefixes.

In general, don't set up any accounts named "test" or "admin" - in mail or elsewhere (judging from my firewall logs, those are the most heavily targeted). And use strong passwords ALWAYS.

• 13. Don't use JPEG for images on your website. It's a lossy format: that means the image will look worse than if you use a lossless format, such as PNG.

13a) That said, JPEG and PNG each have their place. JPG's can be compressed and typically have a smaller file size then PNG's. Unless you compress a LOT, the average person will not be able to notice a difference. Smaller file size decreases page load times, which is an SEO factor..
( Besides - it's just a nice thing to do, for people surfing your site with smartphones, tablets and limited data plans. Unless you're going to go all out and serve different images for each)
PNG - logos, icons, illustrations, text, images which have a lot of high-contrast, distinct lines, images where you need transparency
JPG - photos where colors blend into each other, where you want to serve the smallest possible image size while maintaining acceptable quality. Edit and compress ONCE - (if you compress multiple times, it WILL look like ****)

• Photoshop is not an industry standard. It creates raster images: pictures made of tiny pixels. Industry experts use vector images, with are made of shapes and algorithms. Unlike raster images, vector images don't look worse when they're resized or viewed on high-DPI displays..

14a) PSD isn't an industry standard, but Photoshop can export to industry-standard formats. Newer versions of Photoshop can have "smart objects", which use vector editing. Again, everything has a purpose .
Raster images - Get blurry if you enlarge them, BUT far outperform vectors for color detail.. Better suited for complex images where you want precise color.
Vector images - Can resize without getting blurry, but do not reproduce color detail as well. Best for illustrations, logos, text graphics... You can do gradients, but it can't match raster's pixel-by-pixel color detail. Not good for effects like drop shadows

 

[Paul]

Additional details:

• Mbps = Megabits per second.
• MB/s = Megabytes per second.

There are 8 bits in a byte, so 50 MB/s = 400 Mbps.

To convert from Mbps to MB/s, divide by 8.
To convert from MB/s to Mbps, multiply by 8.

Your math is incorrect because bytes go in groups of 1024, but bits go in groups of 1000:

• 50 MB/s
• Multiply by 1024: 51200 KB/s
• Multiply by 1024: 52428800 bytes/s
• Multiply by 8: 419430400 bps (8 bits in a byte)
  
• Divide by 1000: 419430.4 kbps
• Divide by 1000: 419.4304 Mbps

Therefore, 50 MB/s is about 419 Mbps.

 

[Paul]

13a) That said, JPEG and PNG each have their place. JPG's can be compressed and typically have a smaller file size then PNG's. Unless you compress a LOT, the average person will not be able to notice a difference. Smaller file size decreases page load times, which is an SEO factor..
( Besides - it's just a nice thing to do, for people surfing your site with smartphones, tablets and limited data plans. Unless you're going to go all out and serve different images for each)
PNG - logos, icons, illustrations, text, images which have a lot of high-contrast, distinct lines, images where you need transparency
JPG - photos where colors blend into each other, where you want to serve the smallest possible image size while maintaining acceptable quality. Edit and compress ONCE - (if you compress multiple times, it WILL look like ****)



14a) PSD isn't an industry standard, but Photoshop can export to industry-standard formats. Newer versions of Photoshop can have "smart objects", which use vector editing. Again, everything has a purpose .
Raster images - Get blurry if you enlarge them, BUT far outperform vectors for color detail.. Better suited for complex images where you want precise color.
Vector images - Can resize without getting blurry, but do not reproduce color detail as well. Best for illustrations, logos, text graphics... You can do gradients, but it can't match raster's pixel-by-pixel color detail. Not good for effects like drop shadows

No, typically properly compressed PNG images are smaller than properly compressed JPEG images. The exception is photography. Web UI elements should be SVG/PNG. WebP ("weppy") is better than both PNG and JPEG in nearly all cases, but it's not supported by older browsers.

Vector images can reproduce detail just as well as a raster image--technically, better. Your typical graphic designer just doesn't know how to work with SVG filters, and most don't bother with sub-pixel detail, which isn't possible with raster images (unless you count playing with DPI/PPI ratios, but that's different). As for color detail, SVG supports more color detail than any common raster format. It doesn't matter, though, because consumer screens can't display them. (Common raster formats implement sRGB, which is 24-bit, or 32-bit with the addition of an alpha transparency byte.)

Additional note: SVG filters allow vector images to produce things other than "shapes", but they are very complex and mathematical; most graphic designers don't know much about them, or at least how to chain them. Inkscape is a great tool for experimenting with them.

 

[Joseph Green]

Your math is incorrect because bytes go in groups of 1024, but bits go in groups of 1000:

• 50 MB/s
• Multiply by 1024: 51200 KB/s
• Multiply by 1024: 52428800 bytes/s
• Multiply by 8: 419430400 bps (8 bits in a byte)
  
• Divide by 1000: 419430.4 kbps
• Divide by 1000: 419.4304 Mbps

Therefore, 50 MB/s is about 419 Mbps.

The only time I have ever seen anyone use 1000 instead of 1024 is in the marketing of hardware capacity to trick consumers.

Google agrees with me.

 

[Paul]

The only time I have ever seen anyone use 1000 instead of 1024 is in the marketing of hardware capacity to trick consumers.

Google agrees with me.

Often calculators just divide by 8 for approximation purposes. Bits, unlike bytes, have no reason to go by exponents of two. They stick to conventional decimal (base-10) math.

According to Wikipedia, bits follow SI conventional prefixes, which is why kbps uses a lowercase "k". Additionally, Wolfram|Alpha agrees with me.

 

[enlytend]

No, typically properly compressed PNG images are smaller than properly compressed JPEG images. The exception is photography.

That's what I was referring to.

I was talking "real world" on the image formats. A lot of potential in other formats, but they won't see widespread use until they're universally supported (hw/sw) AND designers have the motivation to conquer the learning curve and the tools to easily implement.

 

[Paul]

That's what I was referring to.

I was talking "real world" on the image formats. A lot of potential in other formats, but they won't see widespread use until they're universally supported (hw/sw) AND designers have the motivation to conquer the learning curve and the tools to easily implement.

Right, it's a lot harder to make a fancy SVG image than it is to make a fancy JPEG or PNG image.

 

[Joseph Green]

Wolfram|Alpha agrees with me.

And now, Wolfram|Alpha agrees with me. The difference here is megabytes (MB) vs mebibytes (MiB). You used mebibytes.

From Wikipedia:

[A mebibyte] was designed to replace the megabyte used in some computer science contexts to mean 220 bytes, which is similar to the SI definition of the prefix mega (106) but conflicts with it.

 

[Paul]

And now, Wolfram|Alpha agrees with me. The difference here is megabytes (MB) vs mebibytes (MiB). You used mebibytes.

From Wikipedia:

Officially MB goes by 1000s. Try MB to bytes. MiB follows an established standard; for MB, Wolfram|Alpha is following SI conventions.

Additionally, Google contradicts itself. It converts 1 kbps to 1000 bps and 1 Mbps to 1000 kbps.

 

[Paul]

And now, Wolfram|Alpha agrees with me. The difference here is megabytes (MB) vs mebibytes (MiB). You used mebibytes.

From Wikipedia:

Wolfram|Alpha uses SI units without the "i", so you were getting 10^6 bytes in a MB, which isn't conventional. I was talking about conventional megabytes, which, by recent convention, should actually be called mebibytes. Nobody besides Google and Wolfram|Alpha know this, so I usually just use MB/s instead of MiB/s. However, we were technically discussing MiB/s. And I know you meant MiB too because you talked about manufacturers misleading customers by using steps of 1000 instead of 1024.

 

[Joseph Green]

Actually, I did this:

• Begin: 50 MiB/s
  = mebibytes per second.
• Multiply by 1024: 51200 KiB/s
  = "50 mebibytes per second in kibibytes per second" on WolframAlpha
• Multiply by 1024: 52428800 Bps
  = "51200 kibibytes per second in bytes per second" on WolframAlpha.
• Multiply by 8: 419430400 b/s
  = "52428800 bytes per second in bits per second" on WolframAlpha.
• Divide by 1024: 409600 Kib/s
  = "419430400 bits per second in kibibits per second" on WolframAlpha.
• Divide by 1024: 400 Mib/s
  = "409600 kibibits per second in mebibits per second" on WolframAlpha.

Except I didn't use WolframAlpha originally.

Either way, I still learned something new today. Thanks, Paul.

 

[Paul]

Actually, I did this:

• Begin: 50 MiB/s
  = mebibytes per second.
• Multiply by 1024: 51200 KiB/s
  = "50 mebibytes per second in kibibytes per second" on WolframAlpha
• Multiply by 1024: 52428800 Bps
  = "51200 kibibytes per second in bytes per second" on WolframAlpha.
• Multiply by 8: 419430400 b/s
  = "52428800 bytes per second in bits per second" on WolframAlpha.
• Divide by 1024: 409,600 Kib/s
  = "419430400 bits per second in kibibits per second" on WolframAlpha.
• Divide by 1024: 400 Mib/s
  = "409,600 kibibits per second in mebibits per second" on WolframAlpha.

Except I didn't use WolframAlpha originally.

Either way, I still learned something new today. Thanks, Paul.

As I explained in my first post, it's possible to mix different notation styles (Bps, b/s), but it's misleading. I recommend sticking to conventional notation: bps and B/s. If you're talking to a particularly technical crowd, using the IEC prefixes for bytes (KiB, MiB, GiB, etc.) can avoid ambiguities, but more often it just confuses people.

 

[Joseph Green]

As I explained in my first post, it's possible to mix different notation styles (Bps, b/s), but it's misleading. I recommend sticking to conventional notation: bps and B/s. If you're talking to a particularly technical crowd, using the IEC prefixes for bytes (KiB, MiB, GiB, etc.) can avoid ambiguities, but more often it just confuses people.

Tell that to WolframAlpha. All of those came straight from WolframAlpha as a result of my very specific queries.

 

[buyselldoms]

Very interesting stuff.

I had a dedicated server with WHM that got hacked with an FTP server flaw. I never use FTP (FTP or FTPS). I don't allow anonymous users etc. I would recommend anyone with FTP server they do not use to simply not run the service. Somehow the person got access to root from it (and not with the root user/password etc). Funny enough despite keeping up to date etc. it was a hardly used server (thank goodness).. but it has opened my eyes to it.

Some shared cloud hosting is awful. Provoke a 500 error it will show someone elses webmaster email. If you kept loading the HTTP 500 page, a new email will appear each time. Not sure why anyone would do that but might be a dream to spammers.

I agree Google has had much more failures than successes. Froogle, google video etc (well, I am not going to name them all)

Apache is very hit and miss. I am still a fan but there are other options out there such as nginx etc. like you stated. All depends on your intended use. Seen so many benchmarks between transfer of small, medium and large, text files, images, and other binaries - and there is a winner for every candidate. Some performs better for some types and sizes of files than others. All production servers should be tuned and security hardened... if you don't have time for that then find a decent shared hosting provider that will have tuned servers and cover the security for you. Furthermore, utilize CDNs and span hosts to speed up page loading time. I generally think choosing a single one-size-fits-all solution isn't the way to go for any website more than a holding page or personal website.

I assume your gas comparison is retail price of printer ink? Fuel is really cheap.

 

[Paul]

Apache is very hit and miss. I am still a fan but there are other options out there such as nginx etc. like you stated. All depends on your intended use. Seen so many benchmarks between transfer of small, medium and large, text files, images, and other binaries - and there is a winner for every candidate. Some performs better for some types and sizes of files than others. All production servers should be tuned and security hardened... if you don't have time for that then find a decent shared hosting provider that will have tuned servers and cover the security for you. Furthermore, utilize CDNs and span hosts to speed up page loading time. I generally think choosing a single one-size-fits-all solution isn't the way to go for any website more than a holding page or personal website.

It's not performance of a single request that's the issue: it's when you get a lot of requests at around the same time. Apache has a lot of issues handling multiple requests; it's very easy to overload an Apache server.

 

[Rich]

Android's market share is larger than iOS', and is growing. iOS' market share is shrinking.

I was surprised at the numbers, almost 80% market share for android, since launching 4-5 years ago? One can only wonder what Google will do in the domain space.

 

[buyselldoms]

It's not performance of a single request that's the issue: it's when you get a lot of requests at around the same time. Apache has a lot of issues handling multiple requests; it's very easy to overload an Apache server.

That is very true, however, many conquer this problem by sticking it behind varnish, squid or nginx etc. Doesn't exactly solve the problem (especially with a poor configuration), more of a work around, but many find Apache very flexible, some of which not found in other competing products.

 

[buyselldoms]

I was surprised at the numbers, almost 80% market share for android, since launching 4-5 years ago? One can only wonder what Google will do in the domain space.

Considering how much Google uses Android in its business, such a dominant market share is anti-competitive. It strengthens everything from its search engine dominance to advertising dominance. This has to be directly comparable to the whole Microsoft Windows and PC market saga. They should be forced by regulators to start charging a licensing fee for the Android system or be fined heavily.

The mobile phone and tablet market needs seriously looking into it. Too much dominance and little overall choice IMO.

 

[Paul]

That is very true, however, many conquer this problem by sticking it behind varnish, squid or nginx etc. Doesn't exactly solve the problem (especially with a poor configuration), more of a work around, but many find Apache very flexible, some of which not found in other competing products.

Actually, that's not such a bad idea. I've never done that, but it would fix many (not all) of the problems.

Personally, I find Nginx to be more flexible than Apache in many ways once you get used to it. It takes a while to learn the configuration concepts, though, especially if you're coming from Apache. Looking back, Nginx' method is more intuitive.

Considering how much Google uses Android in its business, such a dominant market share is anti-competitive. It strengthens everything from its search engine dominance to advertising dominance. This has to be directly comparable to the whole Microsoft Windows and PC market saga. They should be forced by regulators to start charging a licensing fee for the Android system or be fined heavily.

The mobile phone and tablet market needs seriously looking into it. Too much dominance and little overall choice IMO.

I've never heard that view on it. It's an interesting take. I'm an Android fan, but I can see the logic behind that.

 

[buyselldoms]

The average person doesn't even see Google on the same level as corporate giants like Microsoft and Apple - many people probably feel its a non-profit. I am also an android fan, use Google search etc. but if they were less dominant etc. would someone provide a better product/service? That is the question.

After all, Despite everyone using Google search... integrating Google Video into the search results couldn't stop people using YouTube - which Google later acquired.. nothing they could do would beat them. Google's social networking is awful - even the earlier prototypes. Facebook is a rather awful social networking website.

Before the super-brands, if your popular social networking website was rubbish (myspace), another came along and took the people... it happened long before myspace too. Was also the general concept to websites and services in general. Whereas I think the big companies catalyst some tech to market sooner, I also think they are hindering the quantum-leap in internet innovation. Google destroyed the other old school search engines - no one expected something so advanced. In the last 5 years or so there hasn't really been many new players. In theory as more people get online, and the speed that technology advances, ... it should be something that increases not decreases. I love innovation! haha (Someone tell Google despite driverless cars sounding uber-cool, its far from new technology... same can be said for Google Glass)

 

[Paul]

At the same time, Google did all of that because they demonstrated an unprecedented level on innovation. They continue to do so. I interviewed with Google recently, and in the process learned a bit about their infrastructure--not a lot, but enough to be pretty amazed by how they utilize simple, abundant technologies in ways that nobody has thought to. We use some similar methodologies and implementations here at NamePros; on a much smaller scale, of course.

 

[Erin Nagata]

Thanks for sharing these facts. It is important that you should update your password from time to time for your security and convenience too.