Browser

10 things browser developers need to focus on

From a lack of standards to security issues to erratic plug-in support, today's browsers have their share of problems. Here are some areas that could use improvement.

I recently did a quick review of Firefox 5 here on TechRepublic, which made me realize how frustrating and in flux the whole browser war has been. Firefox was once unbeatable. Then it became bogged down with bloat. Chrome came along and with unmatched speed, stole users away from both IE and Firefox. But now Chrome has its own issues (such as errant processes taking the CPU hostage), so Firefox and IE have become the marginal favorites again.

No matter how you slice it, the browser competition is in a state of constant change and turmoil. If it's not a memory leak, it's a lack of standards, or bloat. Of course, it doesn't have to be that way. I'm going to suggest 10 things developers should focus on to bring all browsers to a level where the decision to be made is not based on which is worse, but which is best.

1: Standards

This is the big one, Elizabeth. Holding a browser to standards anymore seems like giving a cat a bath -- it's just not going to happen. Why this is so, at this point in the Internet/computer evolution, I can't figure out. But each browser seems to want to follow its own set of standards. There is a reason the WC3 exists, and it's not to create tests and standards just to toot its own horn. Browser developers need to take a huge step forward toward standardization so that people don't need more than one browser installed just to make sure they can get all their work done.

2: Memory leaks

I am guilty of leaving browsers open all day. I would love to be able to do this without having to worry that my browser (or a tab within a browser) is going to decide to get hungry and devour every CPU cycle my PC owns. I have plenty of other applications I can leave open all day, but I do much of my work within a browser. These memory leaks cause untold problems and data loss.

3: Performance

Give each of your browsers a go and see how much variation you find in performance. Mine are all over the board and can even vary from task to task. Chrome opens insanely fast and renders crazy fast. Firefox appears to handle tabs much better (at least in version 5). Internet Explorer seems to avoid becoming crippled when certain plugins are used. Browsers, regardless of version, need to enjoy the fastest, most reliable performance possible, as they are quickly becoming of the most widely used tools in both business and home.

4: Rendering

There is such a huge discrepancy in how each browser renders, I'm flummoxed as to how developers have yet to jump on the standardization bandwagon to make sure the playing field is leveled. But it's not just the results of the rendering -- the speed at which browsers render also comes into play. And now that mobile device browsers are becoming a factor, rendering on 3G and 4G speeds as well as form factor becomes an issue. Some browsers cache data before rendering, some browsers render on the fly. Which is better? It depends upon who you ask. I think it would be best to have a rendering standard that would help boost both render time and perfect site layout for all browsers.

5: Bloat

Browsers today are full of themselves -- and everything else. Bloat is a huge issue now. Add-ons have certainly helped, because they place the onus on the end user to install or not install extra features. But even with add-ons, browsers (like everything else) seem to suffer from bloat. Instead of adding new features with every release, make it a priority to fix the features (and/or remove some of the current crop) that already exist.

6: Universal plug-ins

Plug-ins for browsers run the gamut of functionality and usefulness. There are some that (for me) are actually must-haves. One of these is Shareaholic, which lets me quickly tweet a page or post a page to Facebook -- something I do frequently to market my articles and my books. But with each browser, even common plug-ins function differently (and in some cases, don't function at all). For example, Shareaholic on Chrome doesn't always pop up the share window. On Firefox, it never fails. Plug-ins need to have universal functionality and work as expected on each browser. Yes, I realize this is as much on the plug-in developer as it is on the browser developer, so it always comes back to el numero uno!

7: Security

As a consultant, I shouldn't be surprised by the number of problems caused by the lack of security in browsers. Yes, there are add-ons to give the browsers a boost. But the core of browser security should come from the browser itself. End users are not always up on the latest add-ons and do not realize that sometimes a browser's security is only as strong as the configuration they've enabled.

8: Cookies

Cookies cause problems. There is no way around that fact. Security issues, login issues -- you name it and cookies cause it. I believe it's time for browser developers to come up with a different subsystem to handle things that are tasked to cookies. With a more secure, easily managed system of tracking information, browsers will become safer and more reliable. (And when they do develop a new system, I certainly hope they come up with a better name than "cookies.")

9: Browser as "OS"

Since the advent of the Chrome OS, I have had a problem with an OS that seems to be, on the surface, nothing more than a giant rendering engine for a Web browser. The browser-as-OS is a limiting concept that might have had its moment, but that moment has passed. With tablets becoming widespread, the mobile device is going to continue to grow in popularity. If the browser-as-OS format gains traction, mobile providers better step it up and improve speeds across the board. Otherwise, these tablets are going to cause untold headaches. And seeing as how the browser is still such a point of frustration with end users, doesn't it makes sense that a browser-as-OS would make for an even larger headache?

10: For Linux

Okay, I realize this is a point of contention (and most likely, a licensing issue), but why is it so hard to have a Linux distribution and/or browser install with the plug-ins that are ready to use? If it's a licensing issue, just have the agreement pop up when the browser is first run. As if any user is going to say "No" to the Flash Player license. Come on! This is a matter of making things as simple as possible for the end user. Everyone uses Flash -- so make it a no-brainer on Linux.

Around the corner

Will these issues cure the world of its uglier problems? No... not even close. But these improvements could make using a browser a much richer experience. Now, I am not a developer. I never claimed to be. I have no idea if these issues can be resolved. But if they can, I highly recommend they be. If they can't be addressed right now, maybe in the near future. Browsing has become one of the single most important tasks on a computer, for many reasons. Making this experience as secure and efficient as possible should be at the top of the list.

About

Jack Wallen is an award-winning writer for TechRepublic and Linux.com. He’s an avid promoter of open source and the voice of The Android Expert. For more news about Jack Wallen, visit his website getjackd.net.

20 comments
sysop-dr
sysop-dr

OK number 2, I use firefox for all of my reference browsing, like Java apis and c++ reference sites, and that sort of thing. I leave it on 24/7 for weeks and months. I use IE for anything Microsoft and our internal web stuff from IT, again on all the time and never shut down. And I use chrome for searching, news and if I need to find specific formula for coding. Again never shut down. And I do not see the memory leaks of which you are referring. Never. (IE 9 or 8 depending on machine, latest chrome or firefox, only restart them for updates. 3 performance, I recommend users use different browsers for different things as each has strengths and weaknesses. The nature of the beast no two teams will make the same thing. 10 Linux. OK I have problems with browsers on Linux. Most of the different distros have their own package install for web browsers and you can't seem to install the original browser for Linux without using the onse from your distro. They fail dependencies or the version available is way out of date (Debian and Chromium for instance, Scientific Linux and firefox...) At least with some other things like say Eclipse I can install the complete thing from Eclipse.org and it covers all of the dependencies. But not browsers. And this accelerated numbering will not be kept up by the Linux vendors.

jscott418-22447200638980614791982928182376
jscott418-22447200638980614791982928182376

I have always been for multiple browsers. It gives uses a choice of features and its one reason Firefox became so popular. But in the end its page rendering and plug in's working correctly that makes the web work. When IE was the only real game in town it was easy for web sites to develop for just one browser. Now of course we have at least 4 browsers for them to try and work with. To me we have created a nightmare for web developers because we lack standards. Or at least we don't enforce standards.

xyman
xyman

"Cookies cause problems. There is no way around that fact. Security issues, login issues you name it and cookies cause it. " I know someone else commented on this but ... what exactly are the security issues that cookies cause?

Art_Jeffries
Art_Jeffries

This is the "downloads" section, but no PDF?

techrepublic
techrepublic

Future browsers should continue to support HTML/CSS/JS as they do now and continue to all do it differently - we can't change that. But - they should all get together an develop a small runtime library which they all support - not as a plugin but built in (so it even works on the ipad). That way you can write simple code (compared to HTML/CSS/JS) using the (single) language of your choice (Basic, C, C++ , C#, Java, whatever) and compile it down to something which all the runtimes can interpret and execute in EXACTLY the same way, thus giving the User the exact same experience on every browser on every OS. Something like SilverLight, but without XAML. I can remember back in 1995 when I wrote my first website and thinking to myself "I can wait until browsers start supporting real languages". Well here we are 16 years later and Microsoft has declared the HTML5 and JS are ther way of the future - what a let down. Back then, the VB3 runtime was about 350MB and apart from the graphics I could write a far better User Experience with it than I can with HTML and JS! - and Ive been using them for 16 years!!

jscott418-22447200638980614791982928182376
jscott418-22447200638980614791982928182376

Has developing standards any good if web sites don't follow them? We seem to focus on the wrong problem here. When we had just Internet Explorer things were actually better even though it lacked much in standards. Why? Because Web sites only had to develop their sites to work with one browser. Now we have many browsers all with there own ideals of what standards to support. Opera may be standards compliant, but it certainly has not helped it gain any user base. So if being standard compliant was so great, why is Opera so un attractive? I think most users could care less about anything except that the browser they use displays web sites properly and allows them to do what they need to do. They could care less if that's compliant to standards or not.

ian3880
ian3880

... and it's called the Opera browser. 1: Standards "But each browser seems to want to follow its own set of standards." NOT SO! Opera is arguably the most standards compliant of all browsers. Can YOUR browser of choice pass the Acid 3 test? [ http://acid3.acidtests.org/ ] Even my highly customised version of Opera v11.11 passes with 100%. See also [ http://www.webstandards.org/action/acid3/ ] for further info on what the Acid 3 test actually checks. "There is a reason the WC3 exists, and its not to create tests and standards just to toot its own horn." Opera has a built-in W3C validation tool, so anyone who wants to see if the web source code is correctly written can not only have the page validated, but are shown information about what is wrong and where. "Browser developers need to take a huge step forward toward standardization so that people dont (sic) need more than one browser installed just to make sure they can get all their work done." Yessss ... it's called browser discrimination - and it's usually not the fault of the browsers developers. It's the fault of web developers who are still writing and/or using exclusive code (going back to the browser wars of the 90's) developed for IE and Netscape Navigator (ie Firefox). The irony is, now that Microsoft is attempting to become current standards compliant, that IE has to have JS patches, like Opera has always needed to get around IE-centric code, to make IE9 work on the abovementioned archaic web pages. 2: Memory leaks Opera? Not guilty. Plug-ins? Don't use any. 3: Performance Opera may not be the fastest kid on the block, but it sure is consistently close to the top. 4: Rendering "There is such a huge discrepancy in how each browser renders, Im (sic) flummoxed as to how developers have yet to jump on the standardization bandwagon to make sure the playing field is leveled." Covered by my comments in 1 above. 5: Bloat "Browsers today are full of themselves and everything else. Bloat is a huge issue now." I have to agree. ALL software is guilty. Sadly even Opera is guilty. There are "features" I don't want, need or use, and which can't be removed easily or not at all. You can have add-ons for Opera, but don't need anything more than Flash and a PDF viewer. That said, Opera 11.5 download is 10.3MB or so (Firefox 5 = 13.6MB, IE8= 16.9MB). Caveat: I don't know if any or all of these browsers still have to go back home to download the complete browser. 6: Universal plug-ins "Plug-ins for browsers run the gamut of functionality and usefulness." ... and greatly reduce the security of the browser. The browser is only as secure as the plug-ins, and Firefox has had a few rogue plugins loaded with malware. "Yes, I realize this is as much on the plug-in developer as it is on the browser developer, so it always comes back to el numero uno!" Yes - do your homework. 7: Security "As a consultant, I shouldnt (sic) be surprised by the number of problems caused by the lack of security in browsers." As a consultant you shouldn't be keeping your blinkers on and only looking at IE.x and Firefox.x - two of the least secure browsers around. Opera was designed in the 90's from the ground up to be secure - and still is. "But the core of browser security should come from the browser itself. End users are not always up on the latest add-ons and do not realize that sometimes a browsers security is only as strong as the configuration theyve enabled." True - but out of the box ANY browser should be in it's most secure configuration - a lot of Mums and Dads out there wouldn't have a clue as to how to enable security configurations. Yeah, OK, they ask their 5 yo children to do it! ;-O 8: Cookies "Cookies cause problems. There is no way around that fact. Security issues, login issues you name it and cookies cause it. I believe its time for browser developers to come up with a different subsystem to handle things that are tasked to cookies. With a more secure, easily managed system of tracking information, browsers will become safer and more reliable." You can, of course, enable/disable cookies on a site by site basis in most browsers, I believe, but once again (and I'm agreeing here) end users shouldn't have to do this, and (probably) 99% of users out there don't even know what cookies are, what they do, and they don't care anyway. Yes, this is one area the browser developers have to get their collective heads around and do better. 9: Browser as OS "Otherwise, these tablets are going to cause untold headaches." Already are. I have to rely on 3G for internet connectivity, and my local cell is already exhibiting infrastructure overload - less than 200kb/s early evenings, but >2.5Mb/s 1 am to about 5 am, then rapidly slows after that. Two years ago it was rare to drop below 1500kb/s. TO SUMMARISE: As far as I know, Opera is arguably the most standards compliant and secure browser out there. It has a very capable email program built in (not a separate email program like Outlook Express or Thunderbird). Although Opera has jumped on the "plug-ins" and "widgets" bandwagon, it works perfectly well out of the box, without either. Yes, you still have to install Flash like everyone else. As installed, it has the inbuilt ability to send web pages to W3C for validation. It can access Opera's Development Tools and analysis tools (Dragonfly). Lastly, Opera developers respond very quickly to reported security issues, and usually a patch (a minor point upgrade) is available within days. Providing those issues are within the browser itself, of course - they can do little about third party security issues (ie Flash). I initially chose Opera in the 90's after doing a lot of research into the other two - basically IE and Netscape. My choice is not about whether it "looks pretty" or is the most popular or other whimsical stuff. It's about functionality, customisation ability, standards compliance and security, SECURITY, S E C U R I T Y. About twice a year I still research and check other browsers for speed and rendering capability or whether there are any worthwhile new 'features', etc., etc. I still believe Opera to be the most standards compliant and the most secure browser available today.

qwer007
qwer007

yeah ,you know what i care about?i think that the most important problem is memory leaks it is a difficult problem which exist in most browsers.yesterday i tried to use avant browser .it is easy happy that avant browser use less memory .good job

Justin James
Justin James

"There is a reason the WC3 exists, and it???s not to create tests and standards just to toot its own horn. Browser developers need to take a huge step forward toward standardization so that people don???t need more than one browser installed just to make sure they can get all their work done." The only version of HTML which browsers can provably adhere to or fail to adhere to is HTML5... you know, the version of HTML that is still being refined and tweaked by the W3C, and the test suites are still being written. "2: Memory leaks" This is a Firefox problem, not a Chrome or IE issue. I have left each of those browsers open, often with a few dozens tabs for weeks if not over a month at a time without issue. "Browsers, regardless of version, need to enjoy the fastest, most reliable performance possible, as they are quickly becoming of the most widely used tools in both business and home." Yes, and all browser vendors are working hard at this. Look at IE 9's massive leap in JavaScript performance, or the shift they are all making towards hardware rendering. "There is such a huge discrepancy in how each browser renders, I???m flummoxed as to how developers have yet to jump on the standardization bandwagon to make sure the playing field is leveled." Where do I begin? HTML deliberately does NOT standardize these things. In fact, HTML doesn't even define a standard style for things. If a browser wants to render the "strong" tag with an underline, it's allowed to. Why? Because the spec never wants to become device-specific. What makes sense for your desktop might not make sense for your phone... or your printer, or a vision-impaired person's text-to-speech software, all of which are considered "rendering devices". "5: Bloat" Go use Chrome, problem solved. Here's the way I see it... Chrome is as stripped down as it gets. IE has the bells and whistles (and it shows). Firefox is for folks who want to use IE, but pretend they don't, so they get a bloatware browser and bloat it up some more with add-ons. Incidentally, it's those add-ons which are usually the cause of your memory leaks you are so annoyed with. "6: Universal plug-ins" Not the worst idea. "End users are not always up on the latest add-ons and do not realize that sometimes a browser???s security is only as strong as the configuration they???ve enabled." You do know that the bulk of the security issues come from the plugins, not the browsers, right? Specifically, Flash, Acrobat, and QuickTime. And you do realize that once an application (a plugin is just an application running inside the browser...) has been installed and authorized to run, the browser has nothing to do with whether or not that application does something dumb, right? The browser vendors, all three of them, are doing a good job at improving security. The plugin folks? Not so much. "I believe it???s time for browser developers to come up with a different subsystem to handle things that are tasked to cookies. With a more secure, easily managed system of tracking information, browsers will become safer and more reliable." Do you realize that cookies are a standard? Aren't you all cranked up because browsers need to follow standards better? Yet, when they do, you turn around and complain about them doing that too? And what exactly would you replace cookies with? Cookies are VERY secure. Browsers only report them to the site that placed them in the first place, sites control their expiration, and browsers store the data in locations that are very locked down. What's the problem? How many security exploits are the result of cookie handling? "9: Browser as ???OS???" Google is the only vendor doing this right now, but it seems like Microsoft is cribbing some of the ideas for Windows 8. Ugh. "10: For Linux" Because Linux distros are overwhelmingly run by people who refuse to bundle anything that doesn't use the GPL, plain and simple. This isn't the browser's fault, and it's not the plugin maker's fault, it's the fault of the Linux distros. Oh, and didn't you earlier say that you want to remove the bloat, and push all of this stuff to plugins? If that's the case, bundling them all up doesn't really help much, does it? J.Ja

erik.langeland
erik.langeland

Many people use their work computer more than their home computer, and they want to use at home what they learn at work. Firefox and Chrome could take a giant leap in popularity if they made it easier rather than harder for businesses to deploy and use their products. While there are benefits to Chrome's frequent release cycle that Firefox has adopted, they should maintain the major.minor version numbering to give IT staff a better idea of how much testing a particular release needs.

apotheon
apotheon

> And I do not see the memory leaks of which you are referring. Do you ever keep more than a dozen tabs open for long periods? I certainly see some "memory leak" problems in Firefox (though technically it's a memory fragmentation issue, apparently). Maybe I just stress my browsers more than you stress yours.

Justin James
Justin James

IE never established a market share strong enough to ignore all other browsers. J.Ja

apotheon
apotheon

It's mostly about privacy. Cookies can be used to track movement across various Websites. The problem is mostly with third-party cookies. Some cookies are poorly applied so that malicious sites can harvest information from them, too -- such as stored passwords.

Parrotlover77
Parrotlover77

Agreed. I know there are those that love JS. More power to them. There are people that love Obj-C too. I don't get it, but hey, whatever floats your boat. Personally, JS to me is just a necessary evil of the web. I can mostly avoid it myself because when I do web development, I let .Net do the heavy lifting. But really, I hate JS with a firey passion. I'm sorry, I just don't like it at all. Give me C# or Java or some other mature object oriented language and I'll be happy. JS? Meh. But that's why I like your common runtime idea. Develop in whatever language you love, it wouldn't matter. That's part of what is so brilliant about .Net.

seeknosy
seeknosy

What you say is true, "most users could [sic] care less about anything except that the browser they use displays web sites properly and allows them to do what they need to do. They could care less if that's compliant to standards or not." A set of standards without widespread adoption is useless, and as you almost imply counter-productive. As you said, Opera is standards compliant, but not widely adopted. However, it's not "standards" themselves that are the goal or the issue. It's COMPLIANCE with standards. If all browser makers would comply with standards, then all browsers WOULD be displaying sites properly and consistently. Then users could choose browsers based on features not tied to standards or site viewing, and still "do what they need to do" no matter what browser or platform they are using.

apotheon
apotheon

> Can YOUR browser of choice pass the Acid 3 test? I think Firefox 4 did, but 5.0 seems to fail with 97 and one of the color blocks coming up gray. Chromium, Surf, and Uzbl all pass with flying colors. > Even my highly customised version of Opera v11.11 passes with 100%. That's interesting, but the Acid 3 test specifies that you should test an uncustomized browser (literally, "default settings"). > Opera has a built-in W3C validation tool, so anyone who wants to see if the web source code is correctly written can not only have the page validated, but are shown information about what is wrong and where. In a well-designed browser, that should be an extension rather than a built-in feature -- unless it is intended to specifically be a development browser, and not a general use browser. > Yessss ... it's called browser discrimination - and it's usually not the fault of the browsers developers. It's the fault of web developers who are still writing and/or using exclusive code (going back to the browser wars of the 90's) developed for IE and Netscape Navigator (ie Firefox). Agreed. Writing code to a specific browser is bad, except in cases where the design in question is never intended to be public. Even then, it might be a little bad. > Opera? Not guilty. Plug-ins? Don't use any. No Flash . . . ? I always have at least one browser with Flash installed, though I try to avoid Flash as much as it is reasonable to do so. > Opera may not be the fastest kid on the block, but it sure is consistently close to the top. Maybe compared to something like IE or Firefox. There are quite a few faster browsers out there, for purposes of general use; they just aren't as popular as the biggest names in the browser market. > ALL software is guilty. Sadly even Opera is guilty. There are "features" I don't want, need or use, and which can't be removed easily or not at all. That's not strictly true. There is software that is not guilty of bloat. Consider slock, for instance -- a screen-locking tool for Unix-like OSes that does nothing but lock the screen and require a password to unlock. There isn't even a password prompt; you just see there's a black screen, and type your password. If you get it right, hitting Enter will unlock the screen. Simple, minimal, and unbloated. It's just not generally very well known software that avoids bloat. I might have doubled the number of people who have heard of slock when I mentioned it in an article here at TR. > That said, Opera 11.5 download is 10.3MB or so (Firefox 5 = 13.6MB, IE8= 16.9MB). By comparison, the complete source archive of Surf is only 320KB, including stuff ignored by the compiler. > ... and greatly reduce the security of the browser. The browser is only as secure as the plug-ins, and Firefox has had a few rogue plugins loaded with malware. (re: plug-ins) I think Jack Wallen might have meant "extensions", to use the Firefox terminology -- which excludes stuff like Acrobat Reader, Flash, and Quicktime. Some extensions actually improve security, such as HTTPS Everywhere and Perspectives. In theory, a good extension system should also allow for the removal of some features from the core software, so that the extension will not (appreciably) reduce the overall security of the browser from its current state, but the ability to skip installing them will certainly improve that security in most cases. > As a consultant you shouldn't be keeping your blinkers on and only looking at IE.x and Firefox.x - two of the least secure browsers around. Opera was designed in the 90's from the ground up to be secure - and still is. Firefox security has definitely suffered as it bloated up. On the other hand, it has at least one distinct security advantage over Opera: verifiable source code. > out of the box ANY browser should be in it's most secure configuration I don't know about "most secure". The most secure configuration is "unusable". The most reasonably secure configuration would be great, though. > As far as I know, Opera is arguably the most standards compliant and secure browser out there. As far as I am aware, that is an unsupportable assertion, because there simply is not enough evidence for Opera's standards compliance relative to some other browsers. > It has a very capable email program built in (not a separate email program like Outlook Express or Thunderbird). I consider that a bad thing. > Although Opera has jumped on the "plug-ins" and "widgets" bandwagon, it works perfectly well out of the box, without either. A good extension system is the one think keeping me from ditching Firefox for Chromium (which I like more than Opera for several reasons). I loathe Firefox, but its extension system allows me a couple of must-have capabilities that are not yet available on Chromium (or anywhere else, I think). On the other hand, even the Firefox extension system is a victim of poor architecture. That's a tale for another day, though. > As installed, it has the inbuilt ability to send web pages to W3C for validation. As mentioned, I'm of the opinion that should be an extension, and not built-in functionality. > Lastly, Opera developers respond very quickly to reported security issues, and usually a patch (a minor point upgrade) is available within days. That's typical of open source browsers. What makes this special about Opera is that it is a closed source browser -- most of which suck on the security patching front. > I still believe Opera to be the most standards compliant and the most secure browser available today. As long as you do not have access to verifiable source, that can never be more than a matter of faith.

Loony Gnoll
Loony Gnoll

Chrome is not perfect, I'd even say Chrome got serious issues like other browsers do. There are memory leaks, profile bloat (a lot of mess in history database files and it's barely controllable by the browser itself), and process-per-tab and page preload overhead itself is ridiculous when it comes to extensions (5-50 megabytes per extension?). These problems are old, I think they've been present in Chrome since first release, and they are not fixed yet. I'm not making this up, these problems are in Chromium issue tracker. Though, no matter how fed up I'm with Chrome, it's probably the best browser available.

apotheon
apotheon

There was a period of time when IE market share was well above 90%, and Microsoft sat on its virtual fourth point of contact believing its dominance of the browser market was complete.

Parrotlover77
Parrotlover77

If JavaScript + HTML5 becomes the dominant desktop development environment, I will quit using computers and live in a cave until the nightmare is over.