Anatomy of a Barracuda

We released the first snapshot of Barracuda (AKA Opera 11.10) yesterday. As you can see, the focus has been mostly on bug fixes so far, although most people seem to be eager to try some new features. …

The raw changelog was actually several times larger than the one that was published for the snapshot. This is fairly common when we produce public changelogs, because not everything is fit for public consumption. There may be things like internal architecture changes and preparations for things that are in the pipeline that can't be made public (yet). We also had to do things like merge lots of crash fixes into a single changelog item, just to keep the changelog down to a more manageable size.

So while the changelog you are seeing seems large (and I doubt that most people will even read most of it), it was even bigger before it was prepared for the public snapshot.

When it comes to new functionality, we have yet to announce what we will be adding to Barracuda. The initial announcement did contain a small hint, but as I mentioned, the first Barracuda snapshot is focused mostly on bug fixes. You can expect more of them in future builds.

New versions do need to strike a balance between new functionality and bug fixes, as it's important to stay competitive in the cutthroat browser market. We can't work exclusively on new functionality at the expense of fixing bugs, or vice versa. We need to do both.

It should be noted that the name of the browser is still "Opera". Barracuda is the code name for version 11.10 (or 11.1x, and future versions will also have various code names associated with them, but the browser is still called Opera.

So while you are waiting for the next announcement, remember to download Barracuda, and report bugs so that we can fix even more of them!

Advertisements

39 thoughts on “Anatomy of a Barracuda

  1. For me the latest versions of Opera have been extremly bug-free.There was a definite dip in quality when 10.50 released, most of that was swept up in 10.60, and 11.0/11.01 have been rock-solid for me, no crashes, no real issues at all with any of the functionality.I'm having some issues with 11.10, but nothing that's unexpected from a pre-alpha. If you want to see what a buggy release looks like, Firefox 4 is looking to be the one! It's where Opera was with 10.50 (and worse IMHO), over a year ago.

  2. While I really like all the features and stuff, I please you to focus on bug fixing as much as possible (and spend the time left on new features), as stability is _the_ criteria for browser choice in my experience……keep up the _great_ work!

  3. It will be interesting t see what other things come from BarracudaWill there be a pic Barracuda with a digram of the fish for new features and bug fixes?

  4. Yes you're right. Latest versions of Opera have been really buggy, so if Barracuda focus mainly on bugfix is probably the best thingHowever… yes I want new (or enhanced: Speed Dial, Link, Extensions) features too ๐Ÿ˜€ Good work

  5. I love bugfix versions 10x better than new-feature versions. In fact, I would go into paroxysms of joy if the entire dev team took a year off working on any new features and just made Opera rock-rock-rock solid. ๐Ÿ˜‰

  6. So while the changelog you are seeing seems large (and I doubt that most people will even read most of it), it was even bigger before it was prepared for the public snapshot.

    I do read all the changelogs. There have been multiple occasions where I thought "hmm, so that was a bug, awesome"

  7. plz focus on the bugfixes!Opera haz enough feature so far… and you can add more of them… …*latter*!

  8. Originally posted by CsendesMark:

    Opera haz enough feature so far

    Then why is everyone requesting more? I'm guessing that even you have posted requests. In fact, yes, one of your most recent threads is a request! Not practicing what we preach, are we? ๐Ÿ˜†

  9. Originally posted by clashcityrocker:

    There was a definite dip in quality when 10.50 released, most of that was swept up in 10.60, and 11.0/11.01 have been rock-solid for me, no crashes, no real issues at all with any of the functionality.

    +1Yep, it was exactly the same for me! Except for some non-working sites like scribd (but which currently works fine), google docs (which also currently almost works fine), otherwise no crashes or instability.Only one big problem for me is huge memory usage, and no, I don't have a lot of RAM. This huge memory leak happens especially after watching flash (YouTube/DailyMotion/etc) videos; I don't know, maybe I open up to 40 videos in a raw, one after the other, and the memory used by Opera grows to 1 GB, and there was 0 GB free RAM available since Opera was at 500 MB memory consumption… And at the end I close all the tabs, only the Speed Dial remains opened, but still around 800 MB of RAM is used by Opera… so to continue in a usable manner, I have to restart Opera… :faint: So memory bug fixes! :yes:

  10. That was good to see the changelog. A lot of bugs fixed. Good work!As an Opera user I expect new implementations for html5, leaving the -o prefix from some css attributes, hw acceleration for canvas (as any other browser on the market). Users can't disable tab stacking and tab bar resizing. Why?Bug fixes that I waiting for: bugs that are presents for a long time. These are: mail icon shows unreadable mail when there is no unreadable mail, browser crashes when I trying css3 animations, buggy shadow when dragging box that has shadow. Will NSL bug ever disappear? Sadly Silverlight and Flash has some problems too.I hope I'm on the right place to ask these.

  11. Not at all. I just hoped that somebody read my opinion. I wanted to write an answer for the poll. It's not black and white, like bug fix or new features. There are several type of bug fixes and several type of features. Thanks for reading and answering!

  12. Originally posted by GreyWyvern:

    I love bugfix versions 10x better than new-feature versions. In fact, I would go into paroxysms of joy if the entire dev team took a year off working on any new features and just made Opera rock-rock-rock solid.

    If that matters anyhow, another "vote" from me to that idea.

  13. Originally posted by DanielHendrycks:

    Thank you for all of those fixes, they are heavily appreciated.

    +1 :happy:

  14. I will add myself to the chorus for bug fixes over new function.While it is understandable that you desire to both deliver new function and fix bugs and you prefer to exclude neither to the benefit of the other, the reality is that in fact bug fix is more important, especially at this point in time when Opera has introduced so many new bugs and not fixed critical old bugs.Forcing function over reliability is lose-lose, but fixing bugs over new function is catch-up and if you don't catch up, you've lost.I would remind Opera Development and Marketing, that:1) Most of us use Opera as a *tool*, we don't benefit from testing it except to find out if old bugs have been fixed, and to avoid catastrophic new bugs.2) As a tool, Opera must perform reasonably well. It can't be overconsumptive of resources, it can't crash, loose or corrupt its config files, and it must be predictible and controllable. Removing the auto-update default was a good step back in the right direction.3) Opera users can not benefit from new function when accummulated(ing) old bugs prevent use of newer versions, and consequently we won't be finding any new bugs for you either, nor retesting, reproducing and re-reporting (yet again) old unfixed bugs carried forward into new versions. 4) While it is also understandable to merge multiple fixes into single changelog items, keep in mind that is precisely what makes it so difficult for Opera users to know if a previously reported bug was fixed, deferred, or ignored. The changlelog is the only bug-report feedback that Opera users get, and when you massage the changelog such that a fixed bug is no longer recognizable, and the new version remains problematic for inexplicable reasons, the user has two choices: retest old reported bugs yet again (necessitating multiple new installs in quarantine), or just stand pat on an old version. Which do you suppose is increasingly the more cost-effective and productive option for long-term Opera users?Bug fix is a prerequisite. Without bug fixes, I can't use old function, let alone new function. But with bug fixes, I can use old Opera function and wait as long as it takes for Opera to deliver new function. And if Opera's old function is working correctly, those who support Opera around the world will not be deluged with support problems and will be free (and more inclined) to upgrade to new Opera versions, and even again recommend Opera to the uninitiated.New speed or new function with bugs remains just bugs. The inevitable process of working through new bugs is only tolerable when old bugs are no longer an impediment to use.

  15. Seems like people prefer balance.I think Opera reached an "almost perfect" state. I don't think Opera needs new features. It has EVERYTHING. Maybe adding a more complete download manager, with threaded downloads for 1GB+ files, and support for automatic downloads from rapidshare/megaupload/whatever…And let's be sincere here. Opera cannot win the browser usage share battle, Chrome already won, Firefox losing a bit every month. But I won't EVER change my main browser. So, for marketing purposes, adding new features "just because"… causes my next comment.Opera as it is, it's perfect. What I currently dislike, it's that Opera became a memory hog and it's unstable. I had to reinstall and re-configure Opera three times since 11.00. Extensions were the biggest issue, crashing like crazy.Believe me – now what people needs is stability. New features are great, now I couldn't live without extensions now, but stability comes first.

  16. Originally posted by Jurgi:

    In the past betas, or even alphas were stable enough for every day use. Lately even first stables are often not stable enough.

    What a sweeping statement! Maybe that's true for you personally, but don't tell us your experience mirrors everyone else's.Speak for yourself, please. I'd rather speak for myself too.And I would say that from my personal experience, you are wrong. I notice that a lot of people praise older versions over new ones, conveniently "forgetting" how the same things were said about those allegedly "more stable" versions.Rose-tinted glasses don't reflect reality.The Opera guys did say that recent versions have been more stable than any previous version, though. And they should know, seeing as they have objective measurements of these things, especially thought the automatic crash logger.

  17. Originally posted by michaelpuermayr:

    stability is _the_ criteria for browser choice

    +1 In the past betas, or even alphas were stable enough for every day use. Lately even first stables are often not stable enough. I understand the pressure of the market and rivalry, but obvious bugs mean uncouraged users. Even my friend, big fan of Opera (since version 5) who recommended Opera to me, switched to Chrome for some time.

  18. Reading the changelog of the Barracuda snapshot was indeed proving wearisome but I always try to do so. Starwind mentioned the issue of feedback for bug reporting users:Originally posted by Starwind:

    The changlelog is the only bug-report feedback that Opera users get,

    He has a point there. Also I wanted to say that I prefer fixing tons of bugs over new features hence I very much appreciated the first Barracuda snapshot. Nevertheless I'm looking forward to the day Opera announces the new features for 11.10.

  19. Originally posted by ouzoWTF:

    If a bug in your program affects 1% of your users, does that mean 1,7 million people are affected?

    No, and numbers doesn't matter. Percentages are important. It can 1 million or 1 person, doesn't matter. 1% is 1%.

  20. @To everyone wanting bugfixes over features.You can not only fix bugs, or use 80% of your time for bug fixing. That's just not sustainable economically. Usually, only 1-5% of all reported bugs are critical in nature (I am not referring here the bugs in snapshots). I am working for a software company and I am programmer myself, so I somewhat understand the workings of releases, features, business value, etc.Fixing bugs has little or no business value. We have bugs in our system not fixed for years, simply because they are not important enough to be fixed. Fixing those bugs would take a reasonably long time and would have zero business value, since with or without them the product is usable. The second bugfixing dilemma is what bugs to fix and what not? Usually a company has to concentrate on the majority of it's users. So if a bug crashes the system, but that condition is happening only for 1% or less of the users, the effort and cost of fixing the bug would be higher than the profit gained from those 1% (this is just an example, companies usually target 80-90% of their total users, the rest 10% is just to put aside).So, in Opera's case the situation is the same, and I am sure that the majority of the important bugs are solved. At least I never had any serious issues with a stable Opera build. Yes, there were small annoyances but that's all. I filed the bugs but I am almost 100% convinced they will be never solved because they are too insignificant, even for me. And I am OK with that.Someone in the first posts said fix all bugs and develop features in the remaining time. From my experience, I can tell you, there is no such thing, as there is no bugfree program.

  21. Originally posted by patkoscsaba:

    So if a bug crashes the system, but that condition is happening only for 1% or less of the users, the effort and cost of fixing the bug would be higher than the profit gained from those 1%

    If a bug in your program affects 1% of your users, does that mean 1,7 million people are affected?

  22. Originally posted by patkoscsaba:

    No, and numbers doesn't matter. Percentages are important. It can 1 million or 1 person, doesn't matter. 1% is 1%.

    For me it does matter. Especially when your product and the products of your competitors are free for the users and a switch is not a difficult thing (not technically meant) for most of them. There is also the fact that you want a higher market share. This is complicated, when the budget for advertising is low and you depend on word-of-mouth recommendation. Nearly two million people who tell "I have a bug that will never get fixed" really does not help you in the process.

  23. Right but they're unlikely to have a crash for the same person that left Opera and those leaving the competitors because of crashes are more likely to go to another high marketshare browser than Opera because of its lower profile.

  24. Originally posted by patkoscsaba:

    Fixing bugs has little or no business value. … So if a bug crashes the system, but that condition is happening only for 1% or less of the users, the effort and cost of fixing the bug would be higher than the profit gained from those 1%

    …Fixing bugs clearly has business value; at some point customers won't pay for buggy products, or use them even when free. But that assessment is also one-sided. It fails to consider why Opera isn't widely supported by ISPs and major e-commerce websites.The cost-benefit analysis ought not to be the cost of bugfix against fees from users (fee's which Opera itself waived a few years back in a bid for greater adoption) but rather the cost of bugfix against benefit from greater end-user, ISP and e-commerce website adoption.While any given Opera end-user may experience only a handful of significant bugs which impairs that their ability to interact with an e-commerce website, OTOH ISPs and e-commerce websites are indirectly (consequentially) impacted by all/most Opera user's inability to use Opera satisfactorily with their websites, and so most major ISPs and e-commerce websites put little or no effort into testing against or supporting Opera. Admittedly, some problems could be eliminated by more testing and standards compliance by ISPs and e-commerce websites, but there is little incentive to better support Opera while its marketshare struggles with adoption.Consider, if Opera fixes a single 1%-exposure bug, while that single fix improves use for only 1% of the Opera community, it improves Opera being used with a far smaller number but larger percentage of e-commerce websites (and it is percentages that are important, as you noted, right?). Fix enough of those 1-percenters such that most Opera end-users can go to most e-commerce websites and perform well, login, manage cookies correctly, not loose/corrupt history, certificates, cache, and render pages correctly including plugins, so they can ultimately "buy" (or at least click on ad-links), then support for Opera end-users becomes more commercially attractive and inturn more e-commerce websites may test and support with a more debugged Opera, which inturn may incent more Opera end-users to upgrade or convert back.However, what has transpired over the last few versions is the opposite trend. Website developers (e.g. ISPs, Amazon, e-Bay, banks, Yahoo, etc.,) perceive that Opera has opted to implement new function (with its concomitant new bugs), new function that is of little use (so far) to a website rather than fix old bugs. Such e-commerce websites have even less incentive to test their website functionality against Opera, because like Opera's end-users, making a website work well with Opera involves maintaining a suite of Opera test versions and tolerating and working around the unfixed old bugs (as well as new bugs). Large commercial websites do the same cost benefit analysis and the cost of supporting Opera increases commensurate with a backlog of unfixed bugs, while the benefit of revenue from Opera end-users remains flat (or decreases) as a percentage of marketshare.While fixing a bug may only improve use for 1% of Opera's end-users, however fixing that same bug has a two-fold leveraged impact across e-commerce websites, wherein they could see: 1) an uptick in completed Opera end-user transactions2) Opera becoming easier to supportObviously more than 1 bug needs to be fixed, and many bugs impact more than the assumed hypothetical 1% of the Opera install base, but the direction needs to be fixing bugs which benefit Opera end-users, their ISPs, and the e-commerce websites they visit versus adding functions which do little to encourage adoption of Opera by ISPs and e-commerce websites.Don't look at the issue from the relationship of Opera to its end-users only, but include the relationship of Opera to ISPs and e-commerce websites. Perhaps Opera development (bug fix and/or new function) ought to be prioritized by what impacts e-commerce. Improving that would seem a win-win-win.Opera's problem has been a lack of adoption, and arguably that can't be due to its price or a lack of function anymore, so perhaps its due to a lack of bug fixing?

  25. Originally posted by Starwind:

    Website developers (e.g. ISPs, Amazon, e-Bay, banks, Yahoo, etc.,) perceive that Opera has opted to implement new function (with its concomitant new bugs), new function that is of little use (so far) to a website rather than fix old bugs.

    This is nothing but speculation on your part. In fact, I find it pretty rude to pretend to represent all these companies. Please stop pretending to speak on behalf of others.

    Opera's problem has been a lack of adoption, and arguably that can't be due to its price or a lack of function anymore, so perhaps its due to a lack of bug fixing?

    That is clearly not the case, considering that Opera isn't more buggy than other browsers.The way Opera differs from the 4 other "main" browsers is that it stands alone. IE has Microsoft, Firefox had(!) Google (and the growth stopped after Google started promoting Chrome instead), Chrome has Google, and Safari has Apple. Opera doesn't have anyone with a monopoly to promote it.

  26. Originally posted by Chirpie:

    This is nothing but speculation on your part. … Please stop pretending to speak on behalf of others.

    One need only look at their support pages or note the number of times one is told to "upgrade to a supported browser".My views are my own. Originally posted by Chirpie:

    considering that Opera isn't more buggy than other browsers.

    I didn't say it was, did I. Originally posted by Chirpie:

    Opera doesn't have anyone with a monopoly to promote it.

    What Opera really needs is a monopoly!!! Golly. Who knew?

  27. Are you saying that both Microsoft and Apple are monopolies? I guess in so far as their operating systems come with a browser (I'm actually surprised Apple hasn't gotten more flack over Safari like MS got with IE). That's not a monopoly though, it's just anti-competitive.Anyways, Opera has to work harder than the other browser vendors to gain users so I don't think that having as many bugs or the like is good for them. They have to beat the competition in as many ways as possible. Doing something as well as others is not good enough if they want to increase their user share.

  28. Originally posted by Astrophizz:

    Are you saying that both Microsoft and Apple are monopolies?

    Microsoft is definitely a monopoly. Apple isn't really a monopoly, but then again, Safari hasn't exactly made it big on the desktop either. But Apple is a huge and very powerful company, so Safari benefits from that.Originally posted by Starwind:

    One need only look at their support pages or note the number of times one is told to "upgrade to a supported browser".

    That still doesn't support your speculation. All it means is that they have decided to support only 2 or 3 browsers, not that they have actively looked at the rest.

  29. That's more of a mixed bag. The observable goal is:1) Aligning more with other browsers that have reached a consensus where behavior was undefined in the HTML4 standard.2) Aligning perfectly with the expanded behavior definitions of HTML5 drafts, and let the nonconformists suffer pressure from coders.3) Hunting and eliminating many small/obscure bugs that crept into the vastly new architecture. For the <1%-class bug reports and site problems, in many cases one fix will repair many broken sites. But focusing on the 50% bugs first (like Facebook) should have fixed some of them already…4) Making it easier to find/fix new bugs, by making large adjustments to the tracing and error-reporting framework; this also improves the web developer tools by providing more info.From what I can grasp of Opera's framework, this task is more straightforward than the work required to update Firefox. It seems also that Opera has sided with the winning extension format, given Mozilla Jetpack will eventually replace XULscript-based extensions. But we really don't have a grasp of what Barracuda has set out to accomplish; this is just preparatory Core work before landing the feature changes that depend on it.

  30. Originally posted by Starwind:

    Most e-commerce websites do not declare support for Opera and many suggest upgrading to a supported browser. That is not speculation, that is observable fact.

    You are shifting your position, which is more than a little disingenuous. I addressed a specific false claim by you, and you responded with what basically amounts to a Gish Gallop.Your claim was that "website developers perceive that Opera has opted to implement new function rather than fix old bugs." I explained how you were clearly jumping to conclusions based on false premises, and how you should not pretend to speak on behalf of others (like these sites).Your response was to shift your position and start talking about all sorts of other things.

Comments are closed.