Browser suckage

[Gokuma]

Currently I'm using old versions of Opera and Firefox. And I'm wondering if I should update or try another browser. Firefox v1.something is a bloated shitbag that runs like shit and crashes on some sites and I don't expect a newer version to actually be slimmed down and optomized because that never happens with big software like this. Opera v8.something works nice and fast but has gotten horribly incompatible with a lot of sites so I just can't use it much. I think heard complaints about newer versions so the question is should I update that or try something else? This suckness has resulted in me actually using IE for a few things like mapquest. BTW, this is on a 700mhz with 128mb ram. But I'm also curious what could even be functional on a 120mhz with 32mb.

[RightField]

try chrome

[EarthQuake]

So you're bitching because your version of Firefox is two major versions out-of-date? I take it you don't update your anti-viruses or anything like that either?

However, on the 120mhz machine, I would probably use Chrome because of it's drastically smaller memory usage. IE and Firefox tend to be the same in that respect, but Chrome uses less than half the memory.

[Graf Zahl]

So you got a computer that's simply too old and weak to run modern software and have to resort to outdated versions to make them run at all. And then you complain that these old versions don't work properly anymore?

What do you expect? Even (or especially) the WWW doesn't care for people who stubbornly refuse to upgrade their hardware once in a while. How old is this computer? It has to be 8 years when I use my own previous hardware as reference.

[Lüt]

IE6 + Maxthon Classic. The only way to tolerably browse.

Though as this forum's most avid Firefox hater, I will admit v2 was considerably less buggy and incredibly more resource-effecient - only about 2-4 times as slow as IE on average. Just remember to close it every 10-15 minutes if you value free RAM.

[Gokuma]

Chrome tracks you so fuck that and google. I'm now using Maxthon and it's pretty nice. Thanks.

[leileilol]

on a 120MHz machine i'd either use Seamonkey when absolutely necessary (less bloaty than Firefox and ironically more of a suite and using the same renderer) and Offbyone for everything else. IE5/6 for Java sites (and using the old msjava vm for that)

[Bucket]

I think the Off By One browser is pretty cool.

[PhilibusMo]

I know this probably isn't relative to your interests but:

OMG just buy (or build) a new computer! If you can't even run interent browsers any more on that computer then do't you think that it might just be time to upgrade?

[GreyGhost]

What is it with you and new hardware - your job or a fetish?

[Maes]

I'd say go with IE. It's the most lightweight of all browsers on Windows platform because -like it or not- it's embedded in the OS itself and is always active. I've seen 5 explorer windows open on a Pentium 166 MMX with 64 MB RAM, using XP SP3. No idea how Maxthon enhances the user experience, but the less, the better, IMHO, without having to resort to using Lynx ;-)

Firefox: way too heavy when RAM is at a budget, so forget it.

Chrome: at least the first version I tried didn't any different than Firefox; with 5 "process-separate" tabs, it easily skyrocketed to 150 MB of RAM usage.

Safari: Made by Apple :thumbs down: plus you just won't be able to run the bloatware and gazillion services it silently installs.

Off by One: pretty cool, yeah, but a bit too minimal. I guess it's better than running obsolete versions of Netscape or IE though.

So to sum up, the most lightweight self-contained, fully-implemented browser one can readily use is still good old IE.

[PhilibusMo]

GreyGhost said:

What is it with you and new hardware - your job or a fetish?

haha, well it was my job until i got made redundent. stupid recession. It just amazes me how slow and old many forum members computers are when a desktop that is about 10x better could be built for roughly £150. I just don't understand why people insist on struggling on with old and out dated hardware that realy can't take modern software. I will admit however that occasionaly old hardware is pretty cool in a nostalgic kinda way, but only once it is properly old like Comadore 64 or ZX Spectrum sort of old. Not just unable to run fire fox at any speed that wouldn't have to be measured on a calander sort of old

[myk]

Gokuma said:
BTW, this is on a 700mhz with 128mb ram. But I'm also curious what could even be functional on a 120mhz with 32mb.

I have a system that has more RAM (256) but has a similar processor (666 MHz), and FireFox 2 works pretty nicely. It's only slow to open, but once you get it running, it's quick. Right now I have 17 open tabs and everything is fine and dandy. I often just leave a blank window open so I don't have to bother with that initial slowness. It is slow in standard GMail, for example, but GMail's HTML mode solves this. To be honest, I usually use standard mode anyway, because it allows me to choose what account sends a message, and the like, regardless of it being slower.

Maybe your RAM will hurt performance, though. I would try it, nonetheless.

FireFox 2 also detects bad scripts, allowing you to close them so they don't eat memory or cause issues. It also restores all windows and tabs, possibly with partly typed content (such as a wiki entry or forum post) when recovering from a system shutdown where you don't close the bowser, a system crash, or itself crashing.

I even hacked my file associations so Favorites open in FF. You do have to use IE to add a Favorite, though (that's about the only use I give to IE on this system these days).

In the older system I'd stick to IE6, although it does have problems on some webpages, because of how old it is. It is more likely to get an error or display pages improperly.

RightField said:
try chrome

It requires a current OS and a powerful system.

Lüt said:
Just remember to close it every 10-15 minutes if you value free RAM.

How much RAM did you have? I keep my Windows 98 system running for days and FireFox does not really have adverse effects.

One thing I did learn about FireFox (1 and 2) is that one needs to clear the downloads list (the small window showing downloads you made recently) often, because if it gets large it does eat too much RAM. Before I figured that out, I thought the browser got slow easily. Once I noticed it, performance was satisfying.

Otherwise, perhaps your memory management was misconfigured (such as applying inadequate cache sizes), but in any case there are useful memory-freeing and administering programs, like RAM Idle when quick free RAM is a must. I have it installed, but personally haven't needed to use it.

One thing that doesn't work too well in FF is YouTube vids embedded in forum threads. Go to the YouTube page instead of watching it on the forum.

Graf Zahl said:
So you got a computer that's simply too old and weak to run modern software and have to resort to outdated versions to make them run at all. And then you complain that these old versions don't work properly anymore?

Your comment is pretty senseless considering his complaints are aimed at finding software that does work well on his system. Had he just complained and wanted to keep using the stuff, it would have been another matter.

[Maes]

PhilibusMo said:

It just amazes me how slow and old many forum members computers are when a desktop that is about 10x better could be built for roughly £150. I just don't understand why people insist on struggling on with old and out dated hardware that realy can't take modern software.

I can quote many actual real-life scenarios where old and outdated systems are still used and maintained (heck, I've got to repair, maintain and often even *build new "old" systems* from scratch almost everyday).

Yeah, it is true that even if just for office use, one could buy the cheapest mobo, cheapest CPU, and cheapest DDR-2 RAM stick and, assuming the old box, PSU, VGA and monitor are usable, spend less than 150 GBP or 200 Eur for a system that's immensely more powerful than what would be definitively considered "too old".

Yet, people often don't take this step. Often, even those 200 Eur can be a serious budget consideration for some of people, and a *new* pre-built system is generally more expensive than that. Not everyone can sit down and replace a few components, hell, most people aren't even able to install the OS and needed drivers on their own.

Often, "upgrading" means buying a completely new system, because CPU sockets/RAM type/VGA card slots, even HD type change too radically to keep within one generation, let alone where there are more than two or three. Even the PSU may not be viable anymore. It's a lot harder convincing someone to take this step rather than "changing a few parts".

I would have a hell of an easier time convincing my superiors to scrap this nightmarish hell of haphazard office boxes ranging from Pentium I to Celeron 478 with spank-new 775 celerons (the very least), all using the same hardware config, but alas, they barely approve expenses for consumables such as printer ink.

Most importantly, people stick to a system because "it just works" or is secondary (e.g. a friend of mine has an Athlon64-class machine in his studio, while his parents use a Celeron with windows 98). The usual scenario is an old system fulfilling a role such as a "typewriter" or just running some accounting program, and which one day proves "too slow for internet" or for "game xxx", and the owner (usually a computer illiterate) asks if it can be "upgraded just a bit". Good luck with that....

[pritch]

Gokuma I'm running latest Windows XP everything on my ancient laptop, as much out of curiousity as anything else. That means SP3 and the latest pre-release candidate of IE8.

Btw it's a 700mhz P3 with 512mb of slow ram, I forget which.

It actually runs fine, the only issues I think being with the modern internet more than the browser, flash is slow and some pages scroll blockily. But IE8 is definitely faster on it than IE7 was. You might need a little more ram though.

[myk]

Maes said:
I can quote many actual real-life scenarios where old and outdated systems are still used and maintained (heck, I've got to repair, maintain and often even *build new "old" systems* from scratch almost everyday).

And in these very forums various people tend to keep old systems to run DOOM in its original state, whith an OPL3-compatible sound card or the like.

[Gokuma]

Firefox is also quick to overheat and shut off this shitty but newer laptop. I already figured out clearing the download list long ago. So now I'm using Maxthon on my two newer comps and I guess I'll try Offbyone on the ancient one. I think it's unwise at this time to blow a substantial amount of money on something I don't need. I usually keep a good safety buffer of money and theoretically I could pay off my car very soon, but my job doesn't seem entirely secure.

What kind of shitty programming is that when a stupid simple list eats up memory like hell?

[GreyGhost]

Much as I admire the effort that's put into open source software, large collaborative projects like Firefox still tend to prove the old adage "a camel is a horse designed by a committee".

[Graf Zahl]

Gokuma said:

What kind of shitty programming is that when a stupid simple list eats up memory like hell?

It's called 'object oriented programming'. :D
It's sometimes really amazing how some younger programmers can waste memory for nothing because they never had to learn how to write efficient code.

[Gokuma]

I want my money back for my computer science degrees.

[Maes]

myk said:

And in these very forums various people tend to keep old systems to run DOOM in its original state, whith an OPL3-compatible sound card or the like.

As someone mentioned, this would fall into the "real oldskool" category: something sufficiently obsolete to only make sense using for a dedicated purpose/nostaglia.

The real problem is what to do with all those systems not sufficiently oldskool to use as dedicated DOS boxes (especially laptops), especially if they don't have SB-compatible soundcards or would otherwise be hard use with pure DOS (e.g. too large HDs etc.)

They may have even been built with a "designed for Windows XP" logo on them, but with components dating back to late 2000/2001, while not being top-notch even for their era. Such a system can be anything in the Pentium II-early Celeron class, with RAM oscillating from 64 MB to 160 MB, typically PC-100, or, if you are lucky, PC-133 or early DDR-266.

That's the catch: they are "designed" to run OSes from Win98 upwards, but can't realistically keep up with managed code apps (Java, .NET) or bloatware such as Quicktime and Flash. And for some reason, they just stick around for too long because they were used in a menial task that didn't go too far beyond their limitations.

[Csonicgo]

Maes said:

I would have a hell of an easier time convincing my superiors to scrap this nightmarish hell of haphazard office boxes ranging from Pentium I to Celeron 478 with spank-new 775 celerons (the very least), all using the same hardware config, but alas, they barely approve expenses for consumables such as printer ink.

That would be the School system my mother has to deal with. I have to repair countless win98 boxes...

[AndrewB]

Graf Zahl said:

It's called 'object oriented programming'. :D
It's sometimes really amazing how some younger programmers can waste memory for nothing because they never had to learn how to write efficient code.

First of all, OOP is not the problem. Second of all, it's not just younger programmers that mess things up for everyone. The problem is with programmers who don't know how to do things efficiently. For example, some programmers might write a routine to scan through every item in a list (could have 50,000 items) just to find the one item that they need. Often they could implement an index using a binary tree to make it run thousands of times faster, but they don't because they've never heard of a binary tree. This is probably a lot more common with programmers who have no formal training. People learn these "important" things when they go to college, whereas hobbyists often just want to see results ASAP.

Also, I am beginning to hate Firefox quite severely. I have a feeling that what I described above has something to do with this.

[Zaldron]

Graf Zahl said:

It's called 'object oriented programming'. :D
It's sometimes really amazing how some younger programmers can waste memory for nothing because they never had to learn how to write efficient code.

This is the stupidest thing I've read all month.

[fraggle]

I can see how object-oriented programming might lead to inefficient memory usage in some situations, although I think that automatically blaming OOP without any investigation of what is causing the problem is a naïve assumption.

OOP is an abstraction that hides some of the underlying details of what's going on. With almost any abstraction like this, it's always possible that you're not doing things as efficiently as you otherwise would. An example is programming in C vs assembler. A human might always be able to write assembly that is more efficient than the compiler generates.

It's the same with OOP. It's possible to do things that are inefficient because you're ignoring the underlying details. For example, every time you dynamically allocate memory (in any language), you always need to use some memory to track the memory that you've allocated. In a language like C (or C++) you can allocate a list of "objects" (or structures in C) as a single big lump of memory. The overhead is less than if you'd allocated each "object" separately.

In a high language like Java or Python it's not actually possible to allocate the objects together, so you can never make this saving. Even in C++ (which Firefox is written in) it's sometimes more desirable or easier to do it this way.

That said, I hardly think that the Firefox developers are inexperienced programmers and it's silly to automatically blame OOP without actually doing any investigation of what's going on. Indeed, naïve attempts to make code "more efficient" by guessing at the cause of the problem without actually doing any profiling to investigate bottlenecks is the worst thing that you can possibly do (and sadly, a very common approach). Furthermore, for a project the size of Firefox, some degree of structuring, like that provided by OOP, is essential for managing the project.

[Graf Zahl]

Zaldron said:

This is the stupidest thing I've read all month.

Are you working in software development? Have you seen some of these people's code? I have.

fraggle said:

I can see how object-oriented programming might lead to inefficient memory usage in some situations, although I think that automatically blaming OOP without any investigation of what is causing the problem is a naïve assumption.

Of course OOP by itself is not to blame. I use OOP myself but I'm also aware of how the underlying interfaces work.

The problem arises when people never learned anything but OOP so they try to apply it to situations where it's just not suited for. This results in extremely inefficient code. But with today's computers this rarely matters. You get away with bad design for small tools because nobody cares if they waste a few megabytes of RAM while running. Apply the same development style to larger programs and you get the kind of bloat which many modern programs have become in recent years.

In a high language like Java or Python it's not actually possible to allocate the objects together, so you can never make this saving. Even in C++ (which Firefox is written in) it's sometimes more desirable or easier to do it this way.

Sometimes yes, sometimes no. This is one of the reasons why I have developed a strong dislike for Java. It leaves me no choice but to write inefficient code.

That said, I hardly think that the Firefox developers are inexperienced programmers and it's silly to automatically blame OOP without actually doing any investigation of what's going on. Indeed, naïve attempts to make code "more efficient" by guessing at the cause of the problem without actually doing any profiling to investigate bottlenecks is the worst thing that you can possibly do (and sadly, a very common approach). Furthermore, for a project the size of Firefox, some degree of structuring, like that provided by OOP, is essential for managing the project.

I won't disagree with that. In fact I also know people who still do this kind of 'naive' optimizations and constantly end up in deeper shit than necessary.

But I still don't see where all the memory in Firefox ends up. That thing is a huge piece of bloat so something in there must be very, very inefficiently done. In the end I often have the feeling that the art of good programming is getting lost - which was the whole point of my (not that serious) post.

[AndrewB]

Graf Zahl said:

The problem arises when people never learned anything but OOP so they try to apply it to situations where it's just not suited for. This results in extremely inefficient code.

It would result in code that is extremely difficult to read or maintain. I'm not convinced it would have a significant performance disadvantage, and would not be the cause of most unjustified slowdowns in our everyday software usage.

In the end I often have the feeling that the art of good programming is getting lost - which was the whole point of my (not that serious) post.

To that I would say that computers are no longer just used for writing summary reports for tire sales quantities.

[Zaldron]

Graf Zahl said:

Are you working in software development? Have you seen some of these people's code? I have.

Yes and yes.
Blaming OOP for what it basically is, simply a bunch of non-POD structs, for consistent lack of perceived performance in an application is ridiculous. Good OOP leaves no holes for terrible bugs nor does it represent significant overhead. If you think the mozilla group doesn't know when to use extern template instantiations, disallow copy constructors in large classes or what STL container to use in each situation you're nuts.

Most of OOP's power is strictly confined to compile-time.

Slowdown can only be blamed to the actual problem-solving the engineers used for each component. The elegance of the code is, in the end, debatable; sometimes security concerns, counter-injection procedures, memory consumption or sandboxing/virtualizing were deemed fundamental in processes you consider straightforward.

The problem arises when people never learned anything but OOP so they try to apply it to situations where it's just not suited for. This results in extremely inefficient code.

Provide a real life example. I mean really, OOP simply organizes your code a bit and stops you from compiling blatant errors..."stuff" is still done in a statement by statement basis..what are you talking about? Don't say something obvious like:

class SimpleNumber {blah};

SimpleNumber num = SimpleNumber();
increment_type it = POST_INCREMENT;
SimpleNumber.Increment(it);

[Graf Zahl]

Zaldron said:

Yes and yes.
Blaming OOP for what it basically is, simply a bunch of non-POD structs, for consistent lack of perceived performance in an application is ridiculous. Good OOP leaves no holes for terrible bugs nor does it represent significant overhead.

hHat's what OOP technically is - but I have seen code that took it too far. Not every programming technique is suited to every problem and yet I have seen countless programmers fresh from the university who were utterly incapable of writing good code.

If you think the mozilla group doesn't know when to use extern template instantiations, disallow copy constructors in large classes or what STL container to use in each situation you're nuts.

No, I actually don't believe that. I was just making fun of current programming trends.

Provide a real life example. I mean really, OOP simply organizes your code a bit and stops you from compiling blatant errors..."stuff" is still done in a statement by statement basis..what are you talking about? Don't say something obvious like:

I have some of such code in front of me every day but it's proprietary stuff I cannot disclose. Most of it was written by a guy who left the company I was working for half a year ago. He was a typical OOP slave - sticking rididly to every OOP paradigm available to the nth degree. The problem now is that this code is so heavily split up into small classes - even using access functions to retrieve and set data in totally passive data structures - that it's impossible to follow or debug any of it without getting totally lost. Why did he do it? Because that's the way his teachers told him.

So please don't try to tell me how great OOP can be - I know that myself. But if you have to deal with the negative side on a constant basis I'm sure your views might change quickly as well. Sometimes having less features so you actually have to think about your code is an advantage.

[Maes]

Heh, this thread so much reminds me of The Daily WTF by now.