Contributing Se Builder to the Selenium Project

[Jason Huggins]

Hello, Selenium Developers!

TL;DR
-------------------------------------------------
If the committers will accept it, I'd like to contribute the
IP for Se Builder to become part of the Selenium Project.

https://github.com/sebuilder/se-builder


Cool Story:
-------------------------------------------------
By way of a brief analogy:
Se Builder is to Selenium IDE as WebDriver is to Selenium RC.

Just as WebDriver gave the Selenium project a new lease on life, I
believe Builder sets up a better baseline to build from in order to
allow record/playback for other browsers and platforms. Bringing in Builder,
however, does not mean we're killing IDE (yet). The goal is to chart
the future for record/playback and plan an orderly transition. And
similar to the merger with WebDriver, realistically, I suspect a
transition of around 2 years. There are lots of IDE users out there
that we all still need to support and take care of.

Moar Story:
-------------------------------------------------
Sauce has been working on the Se Builder [1] project for about two
years now. We believe Se Builder makes Selenium script creation
easier and more manageable. And due to its cross-platform HTML+JS+CSS
architecture, Se Builder should be straight forward to cross port to
browsers beyond Firefox.

With the completion and release of Builder's plug-in architecture and
other improvement (see more complete outline below) we think the time
is right to propose moving the IP for Se Builder under the umbrella of
the Selenium Project alongside Selenium IDE.

Our goal in moving Se Builder to the Selenium Project is:
* Engage the community to push Se Builder forward
* Support the community better by making it easier for the growing
Selenium user community to make better, more supportable Selenium /
WebDriver tests
* Dispel any clouds around Se Builder as vendor-ware that may be
holding back potential supporters / contributors from helping Se
Builder realize the potential we see for it

To be clear, Sauce plans to continue supporting the development of Se
Builder. In the short term we plan to:
- Fix any issues keeping Builder from being fully Se IDE compatible
(minus IDE plugin steps)
- Build a set of plugins enabling integration with commonly used services
- Continue improving usability of the interface

And longer term, the Selenium Project can expect Sauce to:
- Port Builder to Google Chrome and mobile devices
- Support the community by bug fixing, plugin development and documentation

Summary of the current state of Se Builder:
1 - Se Builder now has full native support for Selenium 2 /
Webdriver: users can record, edit, and playback Selenium 2 test
scripts. Se Builder uses the same user interface for Selenium 2 /
Webdriver as for Selenium 1, so users can use conveniently use a
mixture of both technologies. Se Builder can also translate between
Selenium versions with reasonable accuracy.

2 - Se Builder can export Selenium 2 scripts into a wide variety of
languages: Java, Ruby, Python, PHP, JavaScript (Node), and C#, with
more to follow. Se Builder also defines a clean JSON-based format
under-the-hood for storing scripts for later editing. These JSON-based
scripts can be played back by a standalone interpreter. This means you
don't have to commit to a particular programming language for
playback.

3 - Support for plugins that can extend Builder's functionality.
Plugins can be installed and managed directly from the user interface.
Currently, two are available:
- Sauce for Selenium Builder [2], which enables playing back
scripts on Sauce OnDemand directly from the Se Builder interface.
- GitHub Integration [3], which supports storing your scripts
directly to GitHub for safe storage, easy collaboration and access by
CI systems.

4 - The Se Builder code-base is now in a fully non-commercial state.
It contains no reference to or special consideration for Sauce's or
any other commercial system.


So, like I said at the top, if you, the Selenium committers, will
accept it, we would like to contribute the IP for Se Builder to become
part of the Selenium Project. I'm happy to answer any questions that
come up.

Cheers,

- Jason Huggins
http://twitter.com/hugs

[1]: http://sebuilder.com/
[2]: https://github.com/saucelabs/sb-sauce-plugin
[3]: https://github.com/Zarkonnen/sb-github-integration

[David Burns]

+ 1 on accepting builder

David

> --
>
>

[Adam Goucher]

Going on out an limb to have Jason light up my skype in about 30 seconds
I am actually -1 on inclusion and +1 for Builder.

And now for the explanation to the somewhat split opinion.

With WebDriver going through the W3C process and the project's stated
goals as being the maintainer of the reference implementation[s] of that
standard having something that is not part of that process doesn't add
to that goal and just confuses things.

That said, I think Builder is really impressive and it should be
considered [rightfully] as the heir apparent to Se-IDE, just not as part
of the project. I'd love someone to 'svn delete trunk/ide' and just say
'if you want a record/playback solution for webdriver go see this
project over here' much the way we say 'if you want a php/go/node/haskel
solution go see those projects'.

I fully recognize that Se 'won' the automation game on the back of
Se-IDE, but I think the combination of the W3C focus and that /nobody/
with the commit bit actually recommends that people use scripts exported
from the IDE including it as a blessed project seems a bit confusing
from an end user perspective. I can't tell you how often I've had to
explain the dichotomy of having Se-IDE as part of the project but they
can't use it effectively if they want robust, maintainable scripts (aka
Page Object-y since that's the best thing we've come up with).

So yes, Builder /should/ replace Se-IDE -- but not in the main project.
Any more than my php-bindings should be.

-adam

[Patrick Lightbody]

Just got a chance to play with Se Builder for the first time in a long time. I think it's got tremendous potential. There are some usability quirks that I think prevent it from being as brain-dead simple as Se IDE currently is. Some of it will probably get worked out by just adding some new developers with new perspective, but some of it may be hard to avoid due to the inherit complexity introduced by WebDriver.

That said, we have historically done a shit job of taking on a new project and communicating it properly to our users. WebDriver, while a fantastic technical success, was not and is still not very properly communicated to our end users. It also resulted in a fragmented set of users (some of which may benefit from Se Buidler). I would like us to do a better job of integration this time around and use this as an opportunity to finally answer the question "What is Selenium?" and give our users a consistent voice once and for all.

So I'm +1, with two caveats:

1) That we put together and vote on an "integration plan" that tries to right some of our past wrongs.

2) That we promptly update Se Builder to render nicely on a Retina display :)

As for Adam's point below, I don't believe that the W3C process should matter here. Ultimately people are looking to "Selenium", however ill-defined it's been lately, as a place where we can give them tools to automate their browsers. A recorder fits right in line there and we would best serve the community by promoting and further enhancing it.

Patrick

> --
>
>

[Jason Huggins]

On Wed, Nov 21, 2012 at 11:51 AM, Adam Goucher <ad...@goucher.ca> wrote:
> With WebDriver going through the W3C process and the project's stated goals
> as being the maintainer of the reference implementation[s] of that standard
> having something that is not part of that process doesn't add to that goal
> and just confuses things.

I don't see how IDE or Builder interferes with the W3C process of
standardizing the WebDriver protocol.

> /nobody/ with the commit bit actually recommends that people use scripts exported from the IDE

Well, call me "nobody", then. :-) For the past year, as I've been
working on automating iPads, iPhones, and Android mobile sites and
apps with WebDriver, I've been wishing for a "Selenium IDE for iPad"
that bang out quick little scripts against a site as rendered in a
real iOS or Android device. I think this is the appropriate place for
record/playback -- as helper app for creating one or two sample
scripts... which then get exported. The problem is when people /don't/
export and use the IDE as their primary and *only* testing tool. So,
Builder doesn't work as a recorder for iPad today, but I think its'
JS/HTML/CSS underpinnings make a good base for one. And an "Se IDE for
iPad" already passes the "I want that!" test -- which is a pretty big
test in my opinion.

- Jason

[Adam Goucher]

So I flash back to SeConf in SF and the audience was polled about who
uses Se-IDE, and the answer was nobody, and I'll go out on the limb that
those are power users.

Its kinda a bait-and-switch to say 'Here! Use this Recorder ... now
export your code. And now you have to rewrite it to be some sort of
maintainable'.

Most of my incoming sales type calls start with 'I'm using the Firefox
plugin and its not solving my problem' -- will Builder address that
problem in the user base? I doubt it. My canned response to this is
'Automation /is/ programming. Forget that Se-IDE exists outside of these
two very narrow purposes listed on the website and let's talk about how
to structure your code to solve your problems' [1]

Or perhaps a different bar needs to be laid out; how many people, with
the commit bit, would alter their usage of components of this thing we
call Selenium to use Se-Builder as a key component of their automation
tasks. Going out on a limb I would say few/none.

-adam

[1] Yes, this is a biased sample since they are only calling because
they are having problems

[Patrick Lightbody]

Whenever I'm writing a test, I almost always start with Selenium IDE. That's not to say the final product can be consumed by it, but it's a very important part of the experience.

Furthermore, Selenium IDE-only usage is definitely something a large segment of our users that I personally have a desire to continue to support.

-1 big time on downplaying IDE. It's very important, even if our WebDriver committers scoff at those who can't write Java/Python/Ruby/etc.

Patrick

> --
>
>

[Jason Huggins]

On Wed, Nov 21, 2012 at 1:14 PM, Adam Goucher <ad...@goucher.ca> wrote:
> So I flash back to SeConf in SF and the audience was polled about who uses
> Se-IDE, and the answer was nobody, and I'll go out on the limb that those
> are power users.

Right, IDE users don't go to conferences. (They also don't read blogs,
or attend meetups, they're a tough group of people to find.. but they
are out there... they at least know how to download software!)

> Its kinda a bait-and-switch to say 'Here! Use this Recorder ... now export
> your code. And now you have to rewrite it to be some sort of maintainable'.

Not bait and switch if we really try to get better at communicating
that the goal is Record/Export/Refactor, not
Record/Playback/RetireFromThinking.

> Most of my incoming sales type calls start with 'I'm using the Firefox
> plugin and its not solving my problem' -- will Builder address that problem
> in the user base? I doubt it. My canned response to this is 'Automation /is/
> programming. Forget that Se-IDE exists outside of these two very narrow
> purposes listed on the website and let's talk about how to structure your
> code to solve your problems' [1]
>
> Or perhaps a different bar needs to be laid out; how many people, with the
> commit bit, would alter their usage of components of this thing we call
> Selenium to use Se-Builder as a key component of their automation tasks.
> Going out on a limb I would say few/none.

Are you counter-proposing that we remove IDE from the project?

- Jason

[Adam Goucher]

On 2012-11-21 2:39 PM, Patrick Lightbody wrote:
> -1 big time on downplaying IDE. It's very important, even if our WebDriver committers scoff at those who can't write Java/Python/Ruby/etc.
>

So. Devil's advocate time... who is active on Selenium other than
WebDriver committers? Who of the existing project committers is going to
support the users of Se-Builder?

*crickets*

The answer is of course the people who build Se-Builder, which do not I
believe (but could be wrong) have the main Se commit bit. So if they are
supporting it, why do they need to be part of Se where no one [existing]
is going to support it?

I think the phrasing of the question might still be wrong. Its not about
the IP being donated to the project. It is 'do we want to merge the Se
project with the Se-Builder project' much like the Se project and
WebDriver project merged.

Even with the potentially better worded question, I am still -1 on it
since users /will/ use it in a manner that leads to broken,
unmaintainable scripts and tears. (Even though that almost guarantees me
more clients.)

(FWIW this is just part of a much larger project governance issue about
the goals of the project, how to add this, what to add, etc.)

-adam

[Adam Goucher]

Absolutely (though I am not sure that its really a counter-proposal but
a rehash of what I've been saying for awhile). Get rid of Se-IDE and
point all those interested at Se-Builder, which is a separate 'friend of
Se' project. Which is the same status as a lot of moving parts of a
modern Se-based automation setup.

-adam

[Patrick Lightbody]

On Nov 21, 2012, at 11:59 AM, Adam Goucher <ad...@goucher.ca> wrote:

> On 2012-11-21 2:39 PM, Patrick Lightbody wrote:
>> -1 big time on downplaying IDE. It's very important, even if our WebDriver committers scoff at those who can't write Java/Python/Ruby/etc.
>>
> So. Devil's advocate time... who is active on Selenium other than WebDriver committers? Who of the existing project committers is going to support the users of Se-Builder?
>
> *crickets*

I don't like where this conversation is going. It sounds a lot like "if you don't commit, you don't have a vote".

I certainly don't commit to WebDriver's source - it's far too advanced for me to properly grok and be productive with. But that doesn't mean I don't still view myself as a contributor to the project. I'm one of the few active voices on the Conservancy leadership team, and I provide (albeit limited) assistance with infrastructure and mailing list conversations.

Furthermore, I don't believe Selenium is simply an empty vessel for those who most recently have committed code to do whatever they like with it. To be clear: I don't believe that is what's happening with the WebDriver committers.

But I do believe that Selenium is more than just the code: it's a brand and a community. And even though my commit bit is a little dusty I don't think my voice should be any less diminished simply because WebDriver's source code has outgrown my skills.

Patrick

[Patrick Lightbody]

In case it's not already clear: I totally (but respectively!) disagree.

I believe that Selenium should be expanding in scope, especially in light of the W3C. I believe our role in the universe is to promote the brand, grow the community, and provide a *collection* of products and tools that various factions of the community benefit from. I think we are underachieving if we view ourselves as simply a set of native wrappers around the W3C spec.

Patrick

> --
>
>

[Leo Laskin]

+1 to Patrick

Patrick

--

[Santiago Suarez Ordoñez]

I'm, as you'd expect, a +1 for converting Se Builder into a community driven project, instead of a company driven one.
Aside from being a "Saucer", what excites me about this move and makes me want to see this happen is:
- Increasing non-Sauce contributions and decision making for turning Builder into the right record&playback tool for our community.
- Getting a truly WebDriver centered recording paradigm to the masses. I can't wait to see how this changes the way new users evolve towards full blown WebDriver tests.

I believe that Selenium should be expanding in scope, especially in light of the W3C. I believe our role in the universe is to promote the brand, grow the community, and provide a *collection* of products and tools that various factions of the community benefit from. I think we are underachieving if we view ourselves as simply a set of native wrappers around the W3C spec.

I *love* this goal and fully agree with Patrick. It's worth pointing out that the reason for which WebDriver even became what we know as Selenium 2 these days is that the project is rather a "brand" than a tool, with the community recognition as its biggest asset. We shouldn't limit the potential that this "brand" has to define what UI test automation will mean in the future.
We all know record&playback is a fundamental piece of the test automation ecosystem. Isn't the Selenium community the best qualified to decide what it should become and how it should be done?

For the IDE vs Builder fanboys (in case those even exist): We're talking future proof here. Let's not forget Builder is actually just Selenium IDE, with most of its Firefox specific XUL UI replaced with the awesome cross-browser web technologies that every single one of us already know.

[Dave Hunt]

+1 from me for Builder being made the official ultimate upgrade for IDE users. Unfortunately (as one of the IDE maintainers) I have had very limited time recently work on ot, and would happily change focus to work on Builder for any time I have available. In particular, I would be interested in modifying the Mozmill tests I have for IDE to work against Builder (unless there are already suitable tests).

I think the important part is how this would be announced to users, and if there were to be an ultimate end to IDE maintenance then we would need to give plenty of notice, and provide as much assistance to migrate off of IDE as possible. This might involve adding splash screens to IDE or even a wizard to migrate scripts to Builder.

Dave

[Luke Inman-Semerau]

-1

I would prefer to include Se-Builder into the docs and wiki and even downloads page, but keep the source in the repo it is currently in.

I believe the source may receive fewer contributions being rolled into the current Se code base. There are staggeringly few contributions to Se-IDE currently. I thinking giving Se-Builder a first class citizen when it comes to sehq/wiki/downlads(both sehq & googlecode) will give more exposure and hopefully more contributions - given we adequately link to the source repo.

If we do pull the source in, I'd want to know who would be 'responsible' for releases? What time frame would they be in? (Same as Se?) Is there a dependency in Se-Builder on Se itself right now? (Who's gonna create the crazy-fun / Rake target(s)?)


-Luke

> --
>
>

[Jim Evans]

I've had to take a few days to digest the thought of this. I think it's no secret how I feel about record/playback tools. For those on this list who may not be familiar with my opinions, I loathe them. If I were to have access to a time machine, the second thing I would use it for (right after visiting Dallas, Texas, 22 November 1963; sue me, I want answers, dammit!) would be to travel to every software company that has ever produced a record/playback feature for writing code and convince them not to do it. I think they produce code that is inelegant and unmaintainable at best, and actually detrimental at worst.

Despite my personal animosity toward such tools, that animosity does not extend to the Selenium project's users, many of whom get their start using record/playback tools. The project can't compete in the marketplace without a record/playback tool as part of the solution, and the current IDE solution leaves things to be desired. If Se Builder can provide a better experience that will benefit our users and allow them at least to generate code to get them started, I'm tentatively +1 on including this one.

Now, I firmly believe we have a documentation and image problem with respect to IDE. We don't educate people on best practices concerning its use, in my opinion. I have more to say on that, but it's more appropriate on a separate thread.

[Leah Klearman]

So, I have neither a) made any significant contributions to Selenium / WebDriver, nor b) used a record-and-playback tool to create automation, but I wonder if it might be possible to extend Se Builder to help people create and use page-objects...kind of a bridge between people who cannot write code on their own and the current best practices of how to use WebDriver.

Regards,

-Leah

--

[Jason Leyba]

My feelings on record-playback are the same as Jim's, except I think Se.Builder should not be part of the project (nor should IDE).  We barely maintain IDE as it is.  I see these as separate projects that should be maintained as such.  This is similar to proxy support - yes proxies are nice to have for browser automation, but it's outside our focus area so we don't include a proxy as part of the project.

--
 
 

[Simon Stewart]

On Wed, Nov 21, 2012 at 12:07 PM, Patrick Lightbody <pat...@lightbody.net> wrote:

 

But I do believe that Selenium is more than just the code: it's a brand and a community. And even though my commit bit is a little dusty I don't think my voice should be any less diminished simply because WebDriver's source code has outgrown my skills.

I totally agree with Patrick on this.

Simon 

[Simon Stewart]

Time to weigh in.

My personal feelings on record/playback tools are broadly in line with Jason and Jim: I can't stand them, and given the option almost always do my dev work TDD.

But this isn't about my personal feelings. This is about how we support our users, and like it or not, a surprisingly vast number of users rely on IDE for their day to day work. The existing IDE is great for Firefox users, but we've had consistent requests for years for an IE version, and mobile browsers are sufficiently different that we should also be considering these. IDE, as it is right now, isn't a great platform for moving forward to support additional browsers. Builder, however, has been designed with this in mind.

So, I'm +1 for including Builder, but would like to echo Patrick's point about putting together and voting on an "integration plan" that tries to right some of our past wrongs around confusing messaging. 

Simon

[Samit]

Hi,

As one of the IDE maintainers, here are my thoughts:-

My vision for the future would be synergy and fusion of code that would lead to a single core for recording and maybe extension (plugins) that can be used not just by IDE and Builder but also by other projects and hopefully in the future become standard like the current WebDriver API. This can in the future become another W3C spec *dreaming*.

The IDE is still used by a large number of users (above 300,000) who use it for their day to day use. I understand that IDE has lot of code that can be hard to understand and very few people very few people are willing to spent the time to get to know it well, but it works. I have seen people who know nothing of programming fiddle around with it for a couple of hours and start being productive. I believe that most programmers would not use IDE or any record and playback tools unless they have no other option, but the world is also full of people who are testers and not programmers. While these testers have limited choices but to use record and playback tools and slowly migrate to being programmers when they outgrow the features provided by these tools. As we know, this often leads to pool of tests stuck in the ancient practices, aka, a mess - totally agree with you Jim. That is another debate in itself. I believe that users should be warned about the disadvantages awaiting them when they go down the path of using record and playback tools and leave it at that. Whether, we as Selenium, want to continue supporting record and playback tools is indeed a question to be answered. I side with Patrick and Simon sentiments that we need to support a flavour of these. If we do not provide one, users are simply going to use another record and playback tool and there will always be someone creating another one to be better than the rest!

With the WebDriver playback in Selenium IDE that was demonstrated in SeConf 2012, it should be able to play back all tests in all supported browsers, and that I think will include mobile and tablet browsers even on iPad, diminishing the gap between the IDE and the builder. I know it has been delayed for a long time for several reasons, but I should be able to commit it in the next couple of weeks.

I think Se Builder has tremendous potential and until it delivers on the promise of being cross-browser, it is not realising that. There would always be some disadvantages that Se Builder would share with IDE. My belief is that, no matter how much we enjoy that everything can be saved into cloud, many of the businesses that currently use IDE would not want their tests, which may contain sensitive stuff like passwords, to be in the cloud. To get access to the local file system in an efficient way usually means having a browser specific wrapper for each browser to provide these services. Which in turn means having several flavours of the Builder that may be a bit of pain to maintain. Being Firefox specific gives IDE the advantage of lesser maintenance. Also the existing plugins are very appreciated by the users.

When I discussed the Se Builder transition with Jason, I mentioned a realistic time line of about two years, maybe a bit more. I had in mind that after a year, discuss it again and tip the scales in favour of IDE or Builder as the outcome may be. Hopefully iPad support would be in Se Builder by that time. :)

Regardless of the record and playback tool being Selenium IDE or Se Builder, I believe that we will not see a large amount of contributions from the community. Due to the nature of these tools, the target audience is not accomplished programmers, but people do not have sufficient programming skills. These users by definition are unable to contribute much and it seems very few people have the motivation to contribute to a tool that they do not benefit from.

As a side note, I am still waiting for permission from sfc to start the unofficial community based discussion site seleniumanswers.com. I was thinking of redirecting people there instead of using the comments on my blog (http://blog.reallysimplethoughts.com) for asking help about Selenium IDE and my plugins and sometimes Selenium RC / WebDriver as well. I believe that it would help users who dislike creating an account with Google to come and discuss stuff as well.

So in the end +1 for Se Builder and +1 for a realistic transition plan and -1 for expecting tons of contributions from the users.

Cheers,
Samit Badle
@samitbadle
http://blog.reallysimplethoughts.com

[Mike Riley]

I have to agree with both Samit and Simon.

As a programmer, I know that I can create better tests by writing code directly to Selenium.

However, the reason Selenium even got in as a test environment at my previous job was because of the IDE.  Our one QA person there was a non-programmer, plus of the people in my developer support group (we wrote the tests), only I knew Java.  So using the IDE as a learning tool was very helpful, and I even still find it useful in providing some of the non-obvious element locators from time to time.  It also allowed our QA person to create some of his own tests, which we were later supposed to code for, but he worked with the newest bugs while we were developing regression tests.

Our program manager was of the belief that you shouldn't even have to write code.  He thought that you should be able to record everything and just have it play back.  We all know that you can't cover all the possible test cases that way, but the fact that Selenium had an IDE that could do that to some degree is what got it in there as a test environment.  Not having it would have meant that it would never have been considered.

As Samit said, a lot of QA people are not developers, so not having something like the IDE of SE Builder will mean you are not servicing a larger portion of your (potential) community.

I would say the only question is: Do you want it to be part of the Selenium project, or a separate one like the various drivers are, which would make it independent of the Selenium delivery schedule?

I could see it as a standalone project, but it should be referenced one the Selenium pages as we do Chrome Driver and IE Driver now.

Mike

[Simon Stewart]

Broadly speaking, the Selenium project falls into three pieces:

1) A record/playback tool

2) A developer-focused API available in a ton of languages

3) A way of running tests remotely

Right now, those pieces are IDE, RC and WebDriver, and the standalone server incorporating grid technology (tm). We know from past experience that having more than one project in each category leads to a certain amount of confusion. We also know that we're fans of allowing carefully managed migrations rather than big bang moves.

My +1 for Builder is to add it as IDE NG. I'd imagine that the hand over process would be very similar to what we've seen with RC/WebDriver: a slow process that takes a long time to complete. In that respect Samit's "two year" window SGTM. What doesn't sound so good to me is putting off making a decision for that long. 

There's not many of us who have delved into the IDE code and have had at least a quick read of the SauceBuilder code too. From that point of view and IME, SauceBuilder looks like a firmer place to build from than IDE. IDE is essentially Core (which is in maintenance mode) with some extra layers added on top (and sunk into Core) to create a Firefox extension. Bits of SauceBuilder are very similar, but it has the advantage of hindsight: that means that it has already made an effort to separate "plan JS" from "browser specific" in a more robust way than IDE does. That's probably just a reflection on the relative ages of the code bases, but if we have the chance for a clean, stable do-over that's already functionally equivalent, I think it'd be a Good Thing to take it.

Simon

--
 
 

[Patrick Lightbody]

So it sounds like the conclusion is for us to accept it, with the caveat that we ask Jason & Samit to propose an integration plan (perhaps based on Samt's email in this thread) that we can all vote on.

Jason - can you take a stab at that?

Patrick

--
 
 

[Simon Stewart]

Ping Jason or Samit.

Simon

--
 
 

[Paul Grandjean]

Although I haven't had time to read through everyone's comments on this thread, it's fascinating that this debate still continues.  It appears to be an element of Selenium as much as the tools themselves.

P---

--
 
 

[Jason Huggins]

> On Sat, Dec 8, 2012 at 12:44 PM, Patrick Lightbody <pat...@lightbody.net>

>> Jason - can you take a stab at that?

Yes, I'll rally the troops and propose an integration plan --
coordinating with Samit, David Starks, folks at Sauce, and whomever
else I need to pull in. Given the holiday schedule crunch, I think a
plan landing in early January is realistic. I'll shoot for (no latter
than) January 11.

- Jason

[Simon Stewart]

Right. Time to make a decision. Here's the plan:

1) We accept the donation of Builder to the project. It sets us up for the future in a way that existing IDE doesn't.

2) Jason and Samit (or their named minions) come up with a migration plan that ultimately ends up with "... and then we delete IDE".

Things to be bourne in mind (pay attention, things to do are here!)

* We should continue to have a single codebase as everything is closely intertwined.

* Jason: Are there any Saucers who should be given the commit bit as part of this?

* I'm aware that the Builder committers would like to be using git. Because "because".
* Migrating to git is "on hold" until Kristian resurfaces or someone else wants to take on the task. I consider it orthogonal to the migration of Builder.
* Me: I'm happy to own migrating the code into our tree and maintaining the Builder history as much as possible.

Decision made. :) Thank you, Sauce Labs!

Simon

- Jason

--

[Patrick Lightbody]

> Decision made. :) Thank you, Sauce Labs!

Woot!

[Patrick Lightbody]

We don't have that locked down yet, but I think it's safe to say that we won't make the switch until concerns like yours are extremely rare or nonexistent.

Patrick

On Jan 15, 2013, at 6:57 AM, Octavian Covalschi <octavian....@gmail.com> wrote:

May I ask when/if Selenium IDE's support will be dropped?  I'm sorry if I missed this answer in the discussion above, which I read most of it.

The are 2 reasons I'm asking:

1. my personal few days experience with Selenium Builder is horrible, starting from usability UI to basic script recording (I tried on newegg.com, didn't work too well, worked fine with IDE though)

2. we're in a process of setting up an infrastructure for our QA team to use Selenium IDE to create scripts, later which will be converted to junit 4 / Webdriver

So, how much time do Selenium IDE users have? Until things start to break.

Thank you in advance.

--
 
 

[Octavian Covalschi]

Thank you.

[Adam Christian]

Hi Octavia,

We would love it if you could give us all the feedback you have on why the experience was bad -- we have had many responses saying the opposite, so I think its a great opportunity for us to make some improvements regarding your work flow and expectations.

Thanks!

Adam

[Alexei Barantsev]

Any updates on the topic?

Have we got a migration plan?

Regards,

-- 

Alexei Barantsev

Software-Testing.Ru

Selenium2.Ru

[Simon Stewart]

Samit and Jason were meant to be working on it. I guess Jason's been a little busy recently.

Simon

--
You received this message because you are subscribed to the Google Groups "Selenium Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to selenium-develo...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Samit]

Hi,

Based on our very first discussion, I expect a realistic transition period of a least two years. I believe that with this time frame, the users of Selenium IDE would have sufficient time to decide on the right way to migrate. I propose that we see how Selenium Builder has progressed and has delivered on its promise after one year and then decide to put Selenium IDE on end of life. Once on end of life, we will still support it (critical bug fixes only) for one more year to help users migrate ... and then we delete Selenium IDE. :)

Cheers,
Samit

[David Burns]

2 years? That seems unnecessarily long. What is the basis for that time frame?

David

--
You received this message because you are subscribed to the Google Groups "Selenium Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to selenium-develo...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
David Burns
Email: david...@theautomatedtester.co.uk
URL: http://www.theautomatedtester.co.uk/

[Simon Stewart]

This sounds far grumpier than I mean it to be. Please read with your best friendly voice :)

So, AFAIC, it's a done deal that IDE will at some point go away. It comes down purely to how many people are involved and the ability to move the codebase forward. IDE is based on Core, which no-one works on these days, or, and this is the more important point, has any interest in working on, and it's tied to Firefox. The people who have been contributing to IDE recently have been tinkering with formatters and upping the max version of Firefox supported. The last substantive change was made in July last year, when Samit introduced a version of sendKeys backed by the atoms.

RC was far more active than that when we added webdriver into the mix. Builder, OTOH, has seen a healthy number of commits this year. I don't think that the difference is to do with the maturity of the products.

So, I'm curious too: why two years? That seems overly long, especially given how much effort is being put into Core.

Simon

[Samit]

Hi,

Is two years a really long time keeping in view that there are more than 300,000 users out there who will need sufficient time to migrate? These are not the most advanced group of users who will be on the cutting edge of technology and I think we need to keep that in mind when we make such decisions. They are going to need time to experiment and see if any new tools are going to be bug free and suitable for them. During my meetings with users, I have come to realise that many users have invested heavily in their customised versions of Selenium IDE through the use of extensions and plugins that would need to be ported to Builder. We were the ones to encourage them to go in that direction. Let me point out that Builder is NOT compatible with the existing user extensions and plugins. As developers we do know the whole IntelliJ v/s Eclipse v/s your favourite IDE issues as well. Once you are used to a particular IDE, switching to another is not an easy task for everyone. This is true for Selenium IDE users as well. Training people to use and develop expertise to use new tools also takes time. Rebuilding their test infrastructure also takes time. Doing this while trying to meet their day to day deadlines would be indeed very difficult. So in my opinion, we should give them sufficient time. Or are we saying that we do not care about the users who made the Selenium project as popular as it is today?

I believe that we should treat Selenium IDE similar to Selenium RC. In fact they go hand in hand and are joined together by their use of Selenium Core. We did not and still do not tell people to completely stop using Selenium RC and tell them that "we've not stopped fixing bugs in RC and it's not completely frozen". Why then should we treat Selenium IDE users any differently?

I quote from Simon's reply on 5th December:

"My +1 for Builder is to add it as IDE NG. I'd imagine that the hand over process would be very similar to what we've seen with RC/WebDriver: a slow process that takes a long time to complete. In that respect Samit's "two year" window SGTM. What doesn't sound so good to me is putting off making a decision for that long."

I assume SGTM is sounds good to me and not Silently giggling to myself :-)

What has changed between December and now to make the two year window suddenly too long?

I think it will also be fine if, like Simon mentioned in "The path to 3.0" thread, we mark Selenium IDE as legacy when Selenium 4.0 is released, when ever that is. In my opinion, it would amount to the same thing I have proposed.

Cheers,
Samit

[Alexei Barantsev]

Hi devs,

Let's start going this road?

Are there blocking issues? It Builder stable eough? Is it feature complete? Can it be officially recommended as a replacement right away? Should we performs source and the build infrastructure integration? Or keep it as a separate project (forked under SeleniumHQ)?

Regards,

-- 

Alexei Barantsev

Software-Testing.Ru

Selenium2.Ru

[Patrick Lightbody]

Wow. Has it really been a year since we last discussed Builder?! Does anyone know if work continued on it? Personally I was always indifferent to Builder vs IDE from a technology perspective. My concern continues to be that we continue to have a large chunk of users who embrace the record-and-playback style of testing but a committers list full of people who scoff at record-and-playback :P This dichotomy concerns me and I'd love for us to rationalize it!

--
You received this message because you are subscribed to the Google Groups "Selenium Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to selenium-develo...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/selenium-developers/d8cc7649-00b8-4f81-b5f2-9a1320ecfa44%40googlegroups.com.

[Christopher Merrill]

On Sun, Feb 9, 2014 at 11:41 AM, Patrick Lightbody <pat...@lightbody.net> wrote:

Personally I was always indifferent to Builder vs IDE from a technology perspective. My concern continues to be that we continue to have a large chunk of users who embrace the record-and-playback style of testing but a committers list full of people who scoff at record-and-playback :P This dichotomy concerns me and I'd love for us to rationalize it!

FWIW, we are currently running a survey of our users. We've been asking them if they prefer a point-n-click environment or text-based scripting environment. 84% prefer a point-n-click environment (of 132 respondents). Since these are people looking at commercial load-testing tools, they are somewhat self-selected as looking for tools that do NOT require a lot of coding.

I don't have any opinion on the Builder discussion, but thought this statistic might be useful for y'all.

Chris