Mozilla (FireFox et al)

[Cyberman]

Is it just me or is the new versioning for Firefox completely senseless? It's gone from v 3/4 to 14 in only what 8/9 months?

I tried to figure out what they were thinking, and it looks as if they were thinking that 'changing the versioning faster would show faster developement'.

Anyone else do a 'hmmmm' trying to figure it out?

If anyone can tell me what improved I would be glad to hear it, I can't tell apart from most of the new 'features' are annying and waste my time.

Cyb

 

[Toasty]

It's gone from v 3/4 to 14 in only what 8/9 months?

Well Firefox 4 was released in March of 2011, so not quite.

I don't really care about the version numbers -- they're just numbers. They label the different releases, and as long as the numbers are different from each other and progressive, they accomplish their purpose IMO. On the other hand, there is also the new release cycle which accompanied the new versioning system. I guess it doesn't really bother me either. IMO it was probably the lesser of two evils. At the time it was introduced, Chrome was very quickly becoming a major player in the browser market, and there was a lot of pressure on Mozilla to keep FF competitive. Their old system of releasing a slew of new features in a new release every year or two wasn't really keeping pace with the ever-increasing expectations (both of performance and conformance) that Chrome evoked. So they opted for less revolutionary, but more frequent releases that would allow them to get features from development to release in a matter of a couple months instead of potentially years.

FF 4's development definitely could have been executed better, and I think the repeated delays surrounding its release (and the lingering bugs when it finally was released) contributed a lot toward the negative reception the new system has received. And the first few releases after FF 4 also caused a lot of add-on breakage, which did not help matters. IMO, they've kind of 'stabilized' the system now to the point that it's a bit more streamlined. New releases are less disruptive, and while they typically aren't that exciting on their own, if you take the progress that's made in a year (which is about how far apart the releases were before) it's still pretty respectable. I'm looking forward to the built-in PDF viewer in a few weeks.

 

[Cyberman]

You mean they are integrating Evince's engine?

Well The Rev system did need a small change (true). However versioning has been well defined for years. Prefered would have been using MINOR version changes. IE 5.1 5.2 5.3 5.4. and leave the big numbers for major. Revisions 5.0.1 5.0.2 are the next one and sub revisions the last. I use lots of revisions and minor versions.

It does matter too me because major versions are likely to be incompatible or so different there is no going back. For example one is writting in C++ and the new one is written in C# or python type changes. Or one had a simple DB and the next one uses SQLite for data storage.

I understand your POV but too me they seem like Microsoft suddenly (A billion major versions and all are incompatible). A different release schedule is great but that's what minor versions are for? I always have a very strict use of versions because I use it to track exactly what is where and when. I guess it's all a matter of perspective (heh).

Cyb

 

[[SS] Starscream]

You mean they are integrating Evince's engine?

Well The Rev system did need a small change (true). However versioning has been well defined for years. Prefered would have been using MINOR version changes. IE 5.1 5.2 5.3 5.4. and leave the big numbers for major. Revisions 5.0.1 5.0.2 are the next one and sub revisions the last. I use lots of revisions and minor versions.

It does matter too me because major versions are likely to be incompatible or so different there is no going back. For example one is writting in C++ and the new one is written in C# or python type changes. Or one had a simple DB and the next one uses SQLite for data storage.

I understand your POV but too me they seem like Microsoft suddenly (A billion major versions and all are incompatible). A different release schedule is great but that's what minor versions are for? I always have a very strict use of versions because I use it to track exactly what is where and when. I guess it's all a matter of perspective (heh).

Cyb

Not to be a pain in the ass, but you did it twice here so I couldn't resist. Please try to learn how and when to use "too" and "to". A good rule of thumb would be if you can replace the word "too" with "also", that would be a good time to use "too". Like: "Joey and I are going to the store, would you like to go (also) too?"


As far as the version numbers go, it doesn't really matter to me as long as we can tell the difference between versions and that improvements are being made. People will always find something to bitch about if you give them enough time. It could be that they aren't doing enough work or that they're changing revisions too (the "also" rule doesn't apply for this usage of "too") often and they're doing too much work. Oh well.

 

[Toasty]

You mean they are integrating Evince's engine?

I hadn't heard that. AFAIK FF's PDF renderer is being implemented from the ground up in Javascript and HTML5.

https://wiki.mozilla.org/Show_PDF_inline

 

[Cyberman]

Not to be a pain in the ass, but you did it twice here so I couldn't resist. Please try to learn how and when to use "too" and "to". A good rule of thumb would be if you can replace the word "too" with "also", that would be a good time to use "too". Like: "Joey and I are going to the store, would you like to go (also) too?"

I've always puzzled over 'too' but that's a different subject all together. You have my humble apologies for 'too' much.

As far as the version numbers go, it doesn't really matter to me as long as we can tell the difference between versions and that improvements are being made. People will always find something to bitch about if you give them enough time. It could be that they aren't doing enough work or that they're changing revisions too (the "also" rule doesn't apply for this usage of "too") often and they're doing too much work. Oh well.

Honestly normally it wouldn't bother me, but it has been excessive in my perspective. Perhaps it might seem like I'm just complaining, however there are some fundamental flaws in mozilla's source that have been 'put aside' instead of being fixed. These have been lingering for well over 5 years. One of many are issues with the way they've been using the main thread for plugins. I'm not sure Mozilla understands Googles intentions as well as they should either, I've nary seen a thing Google has done as benevolent. Version numbers are important (think milestones). In my perspective it's like saying "well we decided to change the meter because we can't change how fast we walk so instead well make the meter 1/10th of what it is now so we can go faster", when in truth they set there mile stones too far originally. It might be an abstract number too most people but it's not too me. Of course as I said it's a matter of perspective. I guess it appears too me as a political solution than actually solving anything. One thing that has bothered me is the abundant use of Javascript to run firefox.

I hadn't heard that. AFAIK FF's PDF renderer is being implemented from the ground up in Javascript and HTML5.
https://wiki.mozilla.org/Show_PDF_inline

Without chrome's dynamic recompiler this would never work out well (LOL).

PDF is mostly a compressed postscript stream (mostly). Well at least the early versions were, I am uncertain about newer versions. I've not messed a whole lot with PDF and postscript for a decade. (postscript was kind of fun, Javascript seemed ... just wrong as if it were some horrible aboration LOL, PDF fundamentals just appeared as Poscript repackaged with some meta data).

Additional thoughts, I am uncertain it is wise to rely on Javascript for something like that (for that matter much of anything I'm not a fan of Javascript although I've used it in PDFs and the like). It does expand the base of people being able to develop for Firefox's code base. I'm a bigger fan of Python for things that many people are now using Javascript for. Oh well.

Cyb

 

[Surkow]

@Cyberman, if you are interested in the features that are currently being tested (like the pdf reader) I'd recommend you to try Firefox Nightly builds (be sure to backup your profile before doing so). I've been running v17.0a1 for some time now and I consider it to be more stable than the official stable builds.

 

[Cyberman]

LOL I have firefox 9 installed (which I think the latest stable emerge for linux is 13.0.1). Mostly this is about "what's the deal with the major revisions".

Which reminds me I need to look up what's needed to make a plugin, I was thinking it would be nifty making an X plugin for remote display in Firefox. Maybe tweak Evince as a PDF viewer as well (hmm might work).

Cyb - anyhow it's a long road sometimes.

A further update, I've been testing firefox and as noted by Toasty they are making a 'built in plugin' here is why.

It appears the Adobe PDF plugin is unstable to a large degree. Most of this appears to be internet access related (IE reading data from the originating site). I know Adobe likes to track things so I suspect part of it is 'a dial home for the latest' and also just a dial home mechanism. IE you loaded a PDF and they are doing idiot stuff (mostly idiot stuff really is a better term). I've noted this as Firefox has died recently several times under XP but the linux version doesn't (as it doesn't use the Adobe plugin). The Adobe plugin likely assumes it should not play nice (IE do a little bit and relinquish some time back).

Partly too blame is how windows handles threading and processes in general I suspect. Just a suspicion but it can't be corroborated so 'not fact'.

Cyb

addendum 2
I did test the mozilla PDF toy. It has some definate flaws:
1: it has no internal rendering instead it uses HTML5 do to the work
2: It is hard to recognize what icons do what (without 'helper' labels which it lacks)
3: it lacks a traditional TEXT based menu (IE if you can't understand the icons you could use a menu to select options)
4: Index or Preview control is lacking, this is a serious flaw since much of what I am reading in PDF form may have 1200 or more pages in it. It may indicate that it is still "GET IT DONE" mode of thinking. Correction the icons are too blame, they are horribly undichipherable and thus hard to follow. I acidentally found it.
5: More On icons, I'm not sure why they choose the icons but you can't switch cleanly between indexed preview and none cleanly. In addition the icons are designed for visual appeal not usability. The next is more egregious they obviously did not use the settings from the browser (IE font scale) or are even clued in on display resolution. I have a high resolution display (2 1920x1200 side by side 24 inch displays) it is a total <censored> to read things if they decide the scale of things based on pixels in there icons and menus. Although the icons would be probably fine at 1024x 768 they aren't at higher resolutions. Although Beta that is actually paramount in any design that is for presenting visual data (IE reading). I'm not whining read this carefully eye strain is serious if you have high rez monitors I've found, you have to be extremely careful otherwise you get lost on your screens trying to find the idiot cursor because some moron decided to make it 'pixel' sized and not 'pitch' sized (pitch refering to the dpi of your display so that it would be sized so it could be spoted easily in all cases). Let me put it this way I can barely see the cursor with what I'm editing right now because of it.
6: it appears they used the TAB icon size for icons (when it would have been better to use bigger uglier buttons for everyone to easily use of course someone would complain about it \mostly people who don't read a lot of pdfs I suspect\). looking at it proportionately it looks kind of odd
7: lack of prefered default settings. PDFs have 3 sets of settings (surprised?) one is the set the people who made the PDF wanted (which I have found are usually not what I want) and then the settings YOU the user wants and lastly the default settings. It appears the PDF only supports the latter (shrug) somewhat annoying. It has no preferences you can set so you can't overide it's defaults. The majority of PDFS have set whole page and either indexed or preview for pages for a side bar.

Positives:
1: it does work
2: I can block it from working which is handy because some websites do not distinguish between a PDF source and a web source (at all) so that was an unexpected bonus of using it
3: the things that are labeled are labeled in a correct font
4: the rendering of the PDF is decent (which I've found to be an issue with most open source variants of PDF viewers because they tend render things as bit map data instead of vector data) it appears to scale without odd pixelation (reminding me of JPEG decompression from using a DCT or something odd like that).

I don't know if anyone will find my review useful but hopefully they will consider that PDFs sometimes are in the thousands of pages long before murderous attempts at rendering an entire PDF (into bitmaps) at one particular resolution or something (which would be ... seriously huge several gigabytes easily).

To show you want I'm talking about here is a typical data sheet I have to page through (medium sized).

Cyb - back to work