[Opensim-dev] Raise minimum .net framework version to 4.0 and mono version to 2.8 (with 2.10 strongly recommended) in 2Q2013
Ilan Tochner
ilan at kitely.com
Wed Jan 30 00:20:10 UTC 2013
Hi Nebadon,
We've installed versions of Mono (that did not appear in official distro
repositories) on CentOS and Ubuntu. You can find one solution for doing
this here:
http://stackoverflow.com/questions/7538757/mono-2-10-5-for-centos-rhel-5
My point was that a person who can install the latest version of OpenSim
(which AFAIK won't be in a distro repository) can also do the same
personal/IT/bureaucracy work for a newer version of Mono as well. It wasn't
that installing Mono will be painless in that environment it is just that
it will be at the same level of difficulty as installing OpenSim itself.
My only goal in this entire discussion is enabling OpenSim devs, and in
this case Justin, to use .NET 4+ functionality when they require it in
order to avoid hacking around not having that functionality in .NET 3.5. I
think that the limited amount of time that this project's developers have
should not be wasted on things that can be avoided by simply requiring
people to install a later version of Mono when they install a later version
of OpenSim. The people who will need to get permission for installing,
and/or need technical help installing, a later version of OpenSim can just
add installing a later version of Mono as a prerequisite.
I've worked in a company where changing production systems required filing
forms in triplicates and waiting for authorization from multiple
departments until you can get things done. Many Linux applications have
dependencies that need to be upgraded. making Mono one of them is no big
deal if we provide the people who actually do the installation with an easy
way to get Mono installed on their version of Linux (precompiled binaries
or a script-based system such as the aforementioned one should suffice if
they are linked to from OpenSim's installation notes).
let's review the scenario where this will become an issue and see if adding
another step to a complicated (bureaucratically and technically) process is
really what's going to stop someone from upgrading to a later version of
OpenSim that requires Mono 2.10.
As I've said before, your time your decision but there is a cost to not
updating the requirements as well. A cost which I think may be bigger for
OpenSim as a whole than the cost of updating.
Cheers,
Ilan Tochner
Co-Founder and CEO
Kitely Ltd.
<ilan at kitely.com>
On Wed, Jan 30, 2013 at 1:29 AM, Nebadon Izumi <nebadon2025 at gmail.com>wrote:
> Ilan,
>
> I think you are greatly underestimating the ease of installing mono on
> various flavors of linux, I have experience trying to get mono working on
> almost every popular branch of linux you can name and I can speak from
> experience when I say its not as simple as just downloading and RPM or
> adding a new repository, especially in corporate and even educational
> environments you can not just download and install software usually without
> getting permission from the IT department or at least a department head
> depending on how things work at that location. May I ask what flavors of
> Linux you have had Mono and OpenSimulator running on? and What benefits do
> you think we would see by forcing everyone on to new versions of Mono? can
> you be more specific on what improvements in mono you would like to see
> used that we can only use on the newer versions of mono? and how exactly
> would it benefit this project, I am usually the one pushing hardest to get
> us to use newer better versions of mono but not at the cost of upending the
> entire community though. Anyway I look forward to hearing how you think
> this would in technical terms improve the project.
>
>
> On Tue, Jan 29, 2013 at 5:58 PM, Ilan Tochner <ilan at kitely.com> wrote:
>
>> It's been a while but yes, I have worked in a multinational corporate
>> environment. This entire discussion isn't relevant for Windows or Mac
>> users, who can easily install an updated version of Mono if they can
>> install an updated version of OpenSim, so we're already dealing with a
>> small subset of users who use Linux desktops or have access to a Linux
>> server. If someone is capable of manually downloading and installing
>> OpenSim on a Linux machine (by themselves or using the IT department) then
>> that person can just as easily download and install the required Mono
>> version (I assume anyone capable of installing OpenSim is capable of using
>> Google to search for a precompiled 2-year old version of Mono for their
>> particular Linux setup). Is there really someone who has the required
>> privileges and technical know how to install OpenSim, a database, set up
>> ports, etc. without being able to install an additional Mono version on the
>> same machine as well?
>>
>> This isn't a philosophical question it's a question of the number of
>> people who will be affected by whichever option we choose to pursue. I feel
>> we're trying to cater to the needs of hypothetical developers and content
>> creators that run Linux and are able to download, form an external website,
>> and install a software component that isn't in their distro repository
>> (OpenSim) but yet aren't able to download another software component that
>> isn't in their disto repository (Mono) but that can just as easily be found
>> on another website (there are even copies on .edu sites).
>>
>>
>> Cheers,
>>
>> Ilan Tochner
>> Co-Founder and CEO
>> Kitely Ltd.
>>
>>
>> On Tue, Jan 29, 2013 at 8:29 PM, Dahlia Trimble <dahliatrimble at gmail.com>wrote:
>>
>>> Not all of the users of OpenSimulator are sysadmins. Have you ever
>>> worked in a corporate environment? Often the computers that people use are
>>> managed by a central IT department and deviating from the long-term
>>> supported IT mandated solution is not allowed. Similar situations exist in
>>> education. OpenSimulator is not necessarily designed for the sole benefit
>>> of a few for-profit grids, in fact, much of the code base has been
>>> contributed by people who use *and develop* it in such restricted
>>> environments. This is true for content development as well.
>>>
>>> Likewise, not all of the contributors have a large R&D budget. For some,
>>> upgrading to the latest and greatest is not an option, in fact, it could
>>> disable other applications they need to use a shared computer for. Some of
>>> these users have contributed major functionality to the code base. As
>>> stewards of the code base, we need to keep the needs of *all* users in
>>> mind. These are some of the reasons these traditions exist.
>>>
>>>
>>>
>>> On Tue, Jan 29, 2013 at 9:59 AM, Ilan Tochner <ilan at kitely.com> wrote:
>>>
>>>> It's nice to want to support old Linux versions but anyone who is going
>>>> to install OpenSim on Linux, which is not something anyone who isn't a
>>>> capable sysadmin will do, can easily set up that machine with a modern
>>>> version of Ubunto where Mono 2.10 is supported (or use some other distro
>>>> and manually install OpenSim from a third-party website). Had we said that
>>>> upgrading to Mono 2.10 would make it harder for non-technical people to run
>>>> OpenSim then there would have been some (small) justification for holding
>>>> back on advancing OpenSim. However, to do so in order to save sysadmins
>>>> from upgrading their outdated distros is putting the needs of the very few
>>>> above those of the great majority of OpenSim users.
>>>>
>>>> Delaying advancement of a software project that is labeled Alpha for
>>>> the stated reason places enterprise-level legacy support requirements on an
>>>> open-source project that is developed and used by people that have nothing
>>>> preventing them from upgrading their systems (Windows and Mac users have no
>>>> problem using the latest Mono versions as it is). No end-user is going to
>>>> be affected by this upgrade, if any of them is using Linux at home (which I
>>>> doubt more than a few dozen are) then they are likely either using Ubunto
>>>> in the first place and/or are capable of downloading and installing Mono
>>>> from a third-party site.
>>>>
>>>> How many people are going to be served by delaying an upgrade to .NET
>>>> 4.0? How many people will have an inferior OpenSim because of that delay?
>>>>
>>>> It's not that we're preventing anyone from using the existing OpenSim
>>>> version. We're just saying that if you want to use the latest version on
>>>> Linux you need to have Mono 2.10 or later installed. If it doesn't come
>>>> with your distro then search for a third-party site that provides it and
>>>> download it from there. People who can't be bothered to doing either one of
>>>> those things can continue using the existing OpenSim version.
>>>>
>>>> Again, let's focus on advancing OpenSim and not on saving some
>>>> sysadmins the few hours it will take them to install and setup Mono and/or
>>>> a new Linux distro on their server(s).
>>>>
>>>> Cheers,
>>>>
>>>> Ilan Tochner
>>>> Co-Founder and CEO
>>>> Kitely Ltd.
>>>>
>>>>
>>>> On Tue, Jan 29, 2013 at 6:16 PM, Mike Chase <
>>>> mike.chase at alternatemetaverse.com> wrote:
>>>>
>>>>> Just curious. Call me crazy and stuff but why are you worrying about
>>>>> ancient distros with LTS for cases where upgrades to Mono are clearly
>>>>> available. And this is to support software that is perpetually alpha?
>>>>> So
>>>>> you are concerned about adopting .NET 4.0 features because someone
>>>>> might be
>>>>> running an ancient version of debian or Ubuntu presumably in some
>>>>> production
>>>>> scenarios using software you've branded as Alpha.
>>>>>
>>>>> Why don't we call OpenSim what it is. A research project. People
>>>>> have
>>>>> taken and with considerable effort doe some hardening to that
>>>>> sufficient to
>>>>> run a production grid. But it is what it is.
>>>>>
>>>>> And sorry Justin I don't meant to jump on you. You're a good guy. You
>>>>> have
>>>>> to deal with the other members of a board drawing lines in the sand
>>>>> left and
>>>>> right that suit themselves and their own business interests. Sorry
>>>>> Melanie,
>>>>> the "it's never gonna happen" comments are so out of place for a board
>>>>> member of an open public project. Really you have no business being
>>>>> in the
>>>>> position you are. But that's what it is as well.
>>>>>
>>>>> Ok enough ranting. If you feel that upgrading to the 4.0 .NET apis
>>>>> would
>>>>> benefit OpenSim as a whole (I do) then do it. Deciding what versions
>>>>> of
>>>>> mono to use and what distribution to use it on are deployment
>>>>> considerations
>>>>> that someone should be considering carefully based on what they want
>>>>> to use
>>>>> the software for. And if they are trying to run anything close to a
>>>>> production service then they need to be aware of the issues involved
>>>>> in the
>>>>> various versions of mono and make their choice based on that.
>>>>>
>>>>> I doubt I get a vote but if I did I'd vote to advance the API version
>>>>> of
>>>>> .NET and pick up the new features therein. Document the dependencies
>>>>> and
>>>>> let people doing deployment sort out the environment that best meets
>>>>> their
>>>>> requirements. My 2 cents.
>>>>>
>>>>> Mike
>>>>>
>>>>> -----Original Message-----
>>>>> From: opensim-dev-bounces at lists.berlios.de
>>>>> [mailto:opensim-dev-bounces at lists.berlios.de] On Behalf Of Justin
>>>>> Clark-Casey
>>>>> Sent: Monday, January 28, 2013 10:42 PM
>>>>> To: opensim-dev at lists.berlios.de
>>>>> Subject: [Opensim-dev] Raise minimum .net framework version to 4.0 and
>>>>> mono
>>>>> version to 2.8 (with 2.10 strongly recommended) in 2Q2013
>>>>>
>>>>> After some further investigation, it turns out that Mono 2.6 does not
>>>>> make
>>>>> the higher parameter Func calls available unless it has been built in a
>>>>> special preview mode. This is not available on at least the Ubuntu
>>>>> Mono
>>>>> package and I suspect most, if not all, of the other distro packages as
>>>>> well.
>>>>>
>>>>> Therefore, the minimum version of Mono that will use them is 2.8 (for
>>>>> which
>>>>> C# 4.0 is the default). Polling the earliest supported release
>>>>> versions of
>>>>> various Linux distros, the situation is
>>>>>
>>>>> Debian 6.0 (squeeze) Mono 2.6.7
>>>>> Ubuntu 8.04 LTS Mono 1.2.6
>>>>> Ubuntu 10.04 LTS Mono 2.4
>>>>> Ubuntu 11.10 Mono 2.10.8.1
>>>>> Ubuntu 12.04 LTS Mono 2.10.8.1
>>>>> openSUSE 11.4 Mono 2.8.2
>>>>> openSUSE 12.1 Mono 2.10.6
>>>>> CentOS 5 no Mono package
>>>>> CentOS 6 no Mono package
>>>>>
>>>>> I see Debian squeeze as the sticking point here. Debian 7.0 (wheezy)
>>>>> will
>>>>> ship Mono 2.10.8.1 and has been frozen since 2012-06-30. Debian does
>>>>> not
>>>>> work to release dates so it's impossible to say when it will be
>>>>> released,
>>>>> though the indications are that it will be in the first half of this
>>>>> year.
>>>>> Even when it is released, Debian squeeze will most probably be
>>>>> supported
>>>>> until early 2014.
>>>>>
>>>>> In light of this, I am going to recommend that we do not update the
>>>>> minimum
>>>>> version of Mono until Debian wheezy is released. We've already
>>>>> effectively
>>>>> been living with this situation for a while so I don't think that a bit
>>>>> longer is going to hurt, though making modInvoke() properly useable is
>>>>> important. If wheezy is not released by the time that OpenSimulator
>>>>> 0.7.6
>>>>> is here, which I anticipate being shortly after Easter, then we can
>>>>> revisit
>>>>> the topic.
>>>>>
>>>>> This means that existing binary packages will continue to be compiled
>>>>> against .NET 3.5 (though ironically the current 0.7.5-RC packages have
>>>>> been
>>>>> compiled to work with Mono >2.8 only, which will be fixed for the final
>>>>> release).
>>>>>
>>>>> When the update occurs, everything will compile and run against Mono
>>>>> 2.8 but
>>>>> Mono 2.10 will be strongly recommended as the Mono 2.8 series has
>>>>> proved
>>>>> considerably buggy in the past.
>>>>>
>>>>> Once the update is made, the target framework will be .net 4.0 rather
>>>>> than
>>>>> .net 3.5. This will allow c# 4.0 language features to be used and will
>>>>> require the Microsoft .NET Framework 4 packages to be installed on
>>>>> Windows.
>>>>>
>>>>> We've already heard arguments both ways so I doubt that it's worth
>>>>> rehashing
>>>>> them. However, I also think this would be a marginal decision so I
>>>>> welcome
>>>>> any new points.
>>>>>
>>>>> --
>>>>> Justin Clark-Casey (justincc)
>>>>> OSVW Consulting
>>>>> http://justincc.org
>>>>> http://twitter.com/justincc
>>>>> _______________________________________________
>>>>> Opensim-dev mailing list
>>>>> Opensim-dev at lists.berlios.de
>>>>> https://lists.berlios.de/mailman/listinfo/opensim-dev
>>>>>
>>>>> _______________________________________________
>>>>> Opensim-dev mailing list
>>>>> Opensim-dev at lists.berlios.de
>>>>> https://lists.berlios.de/mailman/listinfo/opensim-dev
>>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Opensim-dev mailing list
>>>> Opensim-dev at lists.berlios.de
>>>> https://lists.berlios.de/mailman/listinfo/opensim-dev
>>>>
>>>
>>>
>>> _______________________________________________
>>> Opensim-dev mailing list
>>> Opensim-dev at lists.berlios.de
>>> https://lists.berlios.de/mailman/listinfo/opensim-dev
>>>
>>
>>
>> _______________________________________________
>> Opensim-dev mailing list
>> Opensim-dev at lists.berlios.de
>> https://lists.berlios.de/mailman/listinfo/opensim-dev
>>
>
>
>
> --
> Michael Emory Cerquoni (Nebadon Izumi @ OSgrid.org)
> http://www.onikenkon.com
>
>
> _______________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de
> https://lists.berlios.de/mailman/listinfo/opensim-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://opensimulator.org/pipermail/opensim-dev/attachments/20130130/49ba20fa/attachment-0001.html>
More information about the Opensim-dev
mailing list