Log in

View Full Version : First impressions of final version of great, new Pocket PC Web browser NetFront 3.3


Menneisyys
06-04-2006, 07:38 AM
Because I’m traveling and have to use an expensive and slow GPRS/EDGE connection, I still can’t come up with a fully fledged review (and, in addition, comparison to the brand new Opera Mobile), packed with benchmark data and screenshots. I will definitely publish it after I get home on Monday evening. Please do NOT write in telling me "hey, you've forgotten to mention a very important advantage of NetFront" - I know. This is a quick review, not a fully-fledged one. It, still, already answers a lot of questions of advanced Pocket PC Web browser users, who are really interested in what the final, commercial version of NetFront offers to the users - without just blindly retelling you the contents of the Quick Start Guide.

Now that, thanks to axim_wannabe at AximSite, I’ve managed to get hold of the (Japanese version of - the English version will only follow later (http://www.pocketpcmag.com/blogs/index.php?blog=3&title=final_version_of_netfront_3_3_is_out_in_&more=1&c=1&tb=1&pb=1)) brand new, commercial, 3.3 version of NetFront, the great alternative Web browser for the Pocket PC, I immediately jumped on it to find out in what ways it is different from the latest beta (now trial, 1.06D2) version (http://www.pocketpcmag.com/blogs/index.php?blog=3&p=772&more=1&c=1&tb=1&pb=1) and, as far as its full, commercial version plug-ins are concerned, are the Java and Macromedia plug-ins in any way better than in the previous, 3.2 commercial version.

Differences between the trial and the commercial version

The final version is, on the GUI level, only different in one single checkbox from the trial version released over a month ago. I haven’t managed to find any other differences (except for, of course, the deliberate trial version limitations), not even in the engine – the two programs have exactly the same bugs and the same JavaScript compliance, for example.

Differences between version 3.2 and 3.3

The differences between the two major versions are, of course, much more pronounced than between a one-month-old trial and the final, commercial version. They are as follows:

What is better than before?

* Flawless VGA support (3.2, even when forced into VGA mode, is pretty awkward to use and buggy on VGA devices)
* Flawless WM5 support (3.2 is entirely incompatible with WM5 (http://www.pocketpcmag.com/blogs/index.php?blog=3&title=wm5_aamp_potential_netfront_users_netfro&more=1))
* Enhanced JavaScript support (it’s not as good as that of Opera Mobile, though)
* Greatly enhanced loading and rendering speed. Unlike version 3.2, the new version is definitely faster than the built-in pocket/mobile version of Internet Explorer. It is, however, still definitely slower than Opera Mobile.
* Support for copying link addresses to the clipboard – this is a really welcome addition!
* etc... (see my forthcoming review and those of past, Technical Preview NetFront 3.3 versions!)

What has not changed at all?

* The built-in Java Virtual Machine (JVM) is almost the same as before. That is, it is still not able to run, say, Java Midlets (http://www.pocketpcmag.com/blogs/index.php?blog=3&p=644&more=1&c=1&tb=1&pb=1) (also see this article (http://www.pocketpcmag.com/blogs/index.php?blog=3&p=787&more=1&c=1&tb=1&pb=1) on them) and, compared to the best JVM available today, CrEme (http://www.pocketpcmag.com/blogs/index.php?blog=3&p=637&more=1), it’s clearly inferior (for example, it can only run very few and very simple stand-alone applications, while CrEme can run even the most sophisticated, generic-purpose ones like Toonel (http://www.pocketpcmag.com/blogs/menneisyys/ConfigureToonel45.asp))
* The number of maximal tabs (see below), which is still five
* Macromedia (Adobe) Flash support. I’ve retested the, in version 3.2, non-working Flash animations. Unfortunately, they don’t work with the new version, 3.3, either. This means Flash support hasn’t been improved at all - that is, it’s definitely worse than that of the Pocket Internet Explorer with the (free) Macromedia (Adobe) Flash plug-in. Please see this article (http://www.pocketpcmag.com/blogs/menneisyys/052006FlashPlayers.asp) for more information on this question.
* MHT support (single-file Web archives). While it is able to read small MHT files created on a desktop computer, large ones (I’ve tested with a mostly text-only, 6Mbyte file containing some 1000-1500 forum posts) will result in an “out of memory” error, even if you allocate 70-80 Mbytes of RAM for dynamic memory.

What’s still really missing?

* Hardware button configuration (as with Opera Mobile – the latter also lacks this). The excellent, brand new PIE plug-in PIEPlus 2.0 (http://www.pocketpcmag.com/blogs/index.php?blog=3&p=837&more=1&c=1&tb=1&pb=1) is a perfect example of how this should be implemented.
* The ability to hide the vertical scrollbar at least in full-screen mode



The biggest problems with NetFront 3.3

Maximal number of tabs (windows)

Unfortunately, the new version still has a lot of the problems of version 3.2, as has already been briefly listed above. The biggest problem I have with both version 3.2 and 3.3 is the maximal number of tabs you can open. It’s five, which is definitely worse than that of PIE plug-ins (where, depending on the actual memory overhead of the given plug-in and the free RAM memory size, you can even open 15-20 tabs) and with Opera Mobile, where you can, in most cases, safely keep up to about 30 tabs under WM5 and up to about 25 tabs under pre-WM5 operating systems. (In WM5, because of the vastly enhanced process model, Opera will almost never be automatically shut down; that is, keeping open even 25-30 tabs will not cause any problems. If Access sticks to this severe tab number limitation because of the old, really restrictive and pain-in-the-back "32 processes at most" process model, then, they should reconsider this now that WM5 has got rid of the limitation.)

DST bug still present!

Furthermore, the DST-related bug I’ve discovered almost a year ago (http://www.pocketpcmag.com/blogs/index.php?blog=3&p=326&more=1&c=1&tb=1&pb=1) is still not fixed. This will make it impossible to, say, log into sites that issue cookies valid for less than an hour while your hemisphere is using DST (that is, right now, in Summer time, on the northern hemisphere). Strange Access does not seem to monitor user communities and reviews of their own program. In this respect, they are diametrically opposed to the Opera staff – the latter does follow / contribute to user forums.

JavaScript compliance - could be better!

The JavaScript compliance of the final version, as has already been pointed out, isn’t any better than that of the trial version. That is, for example, you won’t be able to access AJAX-based sites (http://www.pocketpcmag.com/blogs/index.php?blog=3&title=ajax_opera_mobile_8_6_and_netfront_3_3_c&more=1&c=1&tb=1&pb=1) and you won’t have for example address autocomplete in Gmail either (unlike with Opera Mobile, where it does work (http://www.pocketpcmag.com/blogs/index.php?blog=3&p=863&more=1&c=1&tb=1&pb=1). Opera Mobile has much better JavaScript support than NetFront). Spoofing the User-Agent by changing the user-agent to ("Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1)") the desktop PIE as described in here (http://www.pocketpcmag.com/blogs/index.php?blog=3&p=796&more=1&c=1&tb=1&pb=1) , of course, won’t help either.

It won’t support advanced features like dynamic HTML forms / explicit form field disabling (http://www.pocketpcmag.com/blogs/index.php?blog=3&p=865&more=1&c=1&tb=1&pb=1) either (again, unlike Opera Mobile).

HTML compliance - could also be better!

While it’s definitely better than PIE/IEM, it still has the severe HTML parsing bugs greatly influencing for example IHT browsing explained here (http://www.pocketpcmag.com/blogs/index.php?blog=3&p=502&more=1). Another case of Access’ not following English language Pocket PC forums, blogs and news (these bugs have been published by me several months ago), it seems.

Memory leaks, Access plug-ins not terminating

AximSite user axim_wannabe has mentioned (http://www.aximsite.com/boards/showthread.php?t=128504) he’s been experiencing memory leaks with NF 3.3. This seems not to be the case as long as you don’t visit pages with a Flash animation or a Java applet. As soon as you do visit a page like this, the given plug-in won’t exit even if you explicitly exit NetFront. The case is the same with the Java plug-in and as has always been the case with previous NetFront versions. The plug-in architecture of NetFront seems to be somewhat flawed, memory- and control-wise and should be scrutinized by Access. (For comparison, the Java/Flash plug-in architecture of PIE isn’t this bad.)

Verdict

As I’ve pointed out here (http://www.pocketpcmag.com/blogs/index.php?blog=3&p=870&more=1&c=1&tb=1&pb=1), it’s very hard to decide between NetFront and Opera Mobile, currently the two best alternative Web browsers (if you don’t want to stick to the slow and standard-incompliant PIE because of the very weak engine). Both have their merits.

In the article above, I didn’t want to compare the two browsers in every respect (there could have been a lot of other areas to compare – for example, proxy support). I think you may want to choose to wait for my “big” Opera Mobile vs. NetFront comparative review. In the meantime, however, I hope you find my modest quick review helpful.