Home » Modeling » UML2 Tools » I'm worried - UML2Tools vs Papyrus
|
| Re: I'm worried - UML2Tools vs Papyrus [message #558578 is a reply to message #558506] |
Mon, 13 September 2010 14:36   |
 Ed Merks
Messages: 33137
Registered: July 2009 |
Senior Member |
|
|
John,
Comments below.
John McCabe wrote:
> As I understand it, UML2Tools and Papyrus, in principle seem to do
> much the same job. Some time ago I found a UML tool plugin for Eclipse
> that got to a certain point then went commercial. It was a long time
> later that I started looking again and found UML2Tools which seemed to
> fit the bill. However, as others have noted, it seems to have stalled
> without providing an easy route to install in Helios. Papyrus on the
> other hand appears to be actively developed but is not as advanced as
> UML2Tools.
>
> This, to me, is the crux of one of the main issues with Open Source
> Development (and the like). Even if someone does make UML2Tools work
> in Helios, it would appear that it can't be guaranteed that any
> further development will take place.
>
The only way to guarantee ongoing open source development is to do it
yourself or to pay someone to do it. The fact that it's open source
enables that possibility.
> To me there seems to be an awful lot of ego involved in software tool
> development these days.
Egos are involved in all aspects of human social lives.
> Too many people thinking they can do something
> better than the last person who tried and go off on their own merry
> way rather than trying to get involved with the last person who tried
> and helping to shape the product more to their needs (or the needs of
> the people they've surveyed about the product).
>
Of course. Not only that, each generation seems determined to repeat the
mistakes of the previous one.
> It seems to me that the only way round this is commercial product
> development!
>
No, it's fundamentally the same ego problem there. Of course the
difference there is that money exchanges hands so you can demand that
you get value for it. Choose just as carefully though. Many commercial
products also go the way of the dodo, and those that don't will be
revved and modernized so you'll be spending your money again and
learning how to use it again...
> In the meantime though; if I wanted to use UML in Eclipse, what is the
> best option?
>
At this point, Papyrus seems the most active and the most likely to
remain that way.
> John
>
Ed Merks
Professional Support: https://www.macromodeling.com/
|
|
|
| Re: I'm worried - UML2Tools vs Papyrus [message #558599 is a reply to message #558578] |
Mon, 13 September 2010 15:51 |
John McCabe
Messages: 228
Registered: July 2009 |
Senior Member |
|
|
On Mon, 13 Sep 2010 07:36:22 -0700, Ed Merks <Ed.Merks@gmail.com>
wrote:
Ed - thanks for your comments. See below.
<..snip..>
>> This, to me, is the crux of one of the main issues with Open Source
>> Development (and the like). Even if someone does make UML2Tools work
>> in Helios, it would appear that it can't be guaranteed that any
>> further development will take place.
>>
>The only way to guarantee ongoing open source development is to do it
>yourself or to pay someone to do it. The fact that it's open source
>enables that possibility.
That's very true. From a commercial viewpoint though (at least, as a
commercial entity trying to use OSS), this requires investment in one
way or another; either you get someone else to do it, in which case
hopefully you're one of many customers so you basically share the cost
of the development, or you do it in house in which case your
cost/benefit analysis shows a potentially less attractive outcome.
>> To me there seems to be an awful lot of ego involved in software tool
>> development these days.
>Egos are involved in all aspects of human social lives.
Well true, but given the emphasis over the 23 years or so I've been a
professional software developer has been strongly on teamwork, it
would have been nice to believe that, somewhere along the line,
someone may have been paying attention :-) It's ironic (to me) that
many OSS projects are aimed at helping teams work more effectively,
yet the developers working on those projects have chosen not to join
an existing team and, as you say, have ended up repeating the mistakes
of previous projects. My own view of UML in Eclipse seems to support
that!
>> Too many people thinking they can do something
>> better than the last person who tried and go off on their own merry
>> way rather than trying to get involved with the last person who tried
>> and helping to shape the product more to their needs (or the needs of
>> the people they've surveyed about the product).
>Of course. Not only that, each generation seems determined to repeat the
>mistakes of the previous one.
>> It seems to me that the only way round this is commercial product
>> development!
>No, it's fundamentally the same ego problem there. Of course the
>difference there is that money exchanges hands so you can demand that
>you get value for it.
I'm not convinced it's fundamentally the same ego problem. From a
commercial point of view you might believe you can do a better job
than a competitor but that wouldn't necessarily mean you would go
ahead and do it. There are so many variables involved; you need to
know there is a commercial market for another product, you'd have to
make sure you could do it within time and budget constraints yet still
be better than the opposition. It's not just a matter of a guy sitting
in his bedroom deciding "I'll do a better one of those" and going off
on one.
FWIW - programming languages are a case in point there; the
proliferation of different languages these days is, to me, ridiculous.
If, rather than just inventing a new language, some of these talented
people just got involved in working on one of the existing languages
to try to improve it in the particular paradigm they're interested in,
the world would be a much better place :-)
>Choose just as carefully though. Many commercial
>products also go the way of the dodo, and those that don't will be
>revved and modernized so you'll be spending your money again and
>learning how to use it again...
This is true of course, but there are still a number of products that
have been developed in to good, stable commercial products but their
'authors' have decided to abandon them. In some cases, e.g.
OpenWatcom, Delphi, etc they seem to have found their way in to the
public domain for future development and still have some support from
the original development team or at least some of the original
development team!
>> In the meantime though; if I wanted to use UML in Eclipse, what is the
>> best option?
>At this point, Papyrus seems the most active and the most likely to
>remain that way.
Thanks for that. If I'd asked 2 years ago I suspect you could have
replaced Papyrus with UML2Tools :-)
All the best
John
|
|
| | | | |
| Re: I'm worried - UML2Tools vs Papyrus [message #559495 is a reply to message #559000] |
Thu, 16 September 2010 20:28 |
|
Hi all,
My two cents on that subject as I have been involved in the creation of
MDT Papyrus.
1. 2007
In 2007, UMLTools was the only UML2 tool hosted by the Eclipse
Foundation and the Eclipse community hoped it was going to become THE
reference open source UML tool.
Many companies started to use it and to extend it to create modeling
platforms.
In the TOPCASED project, we studied the possibility to rely on GMF
runtime and perhaps on UML2 tools to improve TOPCASED modeling editors.
We faced some issues and submited patches. None of them were analysed by
GMF and by UML2Tools teams during 8 months.
2. EclipseCon 2008
At EclipseCon 2008, there were many participants interested by a UML2
graphical editor hosted by the foundation and we were several different
people to complain that there was no real effort from the Borland guys
to take our patches into account.
There was a "Bird Of a feather" session one evening in which the debate
was strong concerning the strategy of the Borland team, who admitted
that they had no time to consider our patches and had no volonty to
improve UML2Tools for two reasons:
a) UML2Tools was only a test case for GMF : prove that you can
generate 100% of UML2 diagrams with GMF
b) BORLAND had a commercial product based on GMF and they did not want
UML2Tools to provide some competitive functionalities.
So it was then clear that UML2tools was not planned to become a tool of
industrial quality.
If some people were ready to put efforts on an other UML2 tool, the
place was left open.
3. September 2008 : the advent of MDT Papyrus
TOPCASED guys, Papyrus 1 guys (CEA) and Moskitt guys (spanish
initiative) were ready to put efforts to create something of industrial
quality on the basis of UML2 tools.
We started the MDT Papyrus component with the goal to reach industrial
quality level as soon as possible.
4. Why not reusing UML2Tools for MDT Papyrus?
We started reusing UML2Tools. But it was very very hard to maintain as
tha main logics were embedded into the GMFGen templates. We wanted a
modular approach with a "backbone" of services able to provide
transversal services : model explorer, split models, share diagrams with
a single model... all things that were simply not possible with UML2Tools.
After 5 months we decided not to use UM2Tools as foundation as we lost
too much time trying to reuse it...
Sure, you might consider that time has been lost by doing things twice
but they were done in a completely differnt goal and MDT Papyrus aims at
delivering full support from UML2 and industrial quality.
We did some errors by putting two many things in the 0.7.0 release and
spending enough time on validation and bug fixes and performance but we
are going to fix those issues to reach a good quality level in a few months.
5. conclusion
I'm sorry for people who invested a lot on extending UML2Tools and
discover that this is not really maintained but if some of you are
really active then just give us your feedback concerning MDT Papyrus and
you will be given the opportunity to contribute and perhaps become
committer after a few patches.
Welcome.
Hope this context and history helps in understanding this situation,
kind Regards
Raphaël
TOPCASED and MDT Papyrus team
Le 15/09/2010 11:26, John McCabe a écrit :
> On Wed, 15 Sep 2010 03:45:38 -0400, Christian Waniek
> <chris.waniek@gmx.net> wrote:
>
>> I really aggree about this with you. But nevertheless are we using
>> UML2 Tools in our company. So I'm on it to do some develoment for
>> U2T (hoping to become a comitter soon). Because we want to push this
>> project.
>
> Good on you. I wish you the best of luck.
>
>> Although actually it looks like U2T will die, I'm convinced this won't
>> happen because U2T has high potential and there is so much work
>> already done...and the project lead of U2T tries to push this project.
>
>> So I also can't understand why there has to be a new project which
>> actually has the same goal. If you would have invested in u2t as much
>> as there has been to papyrus, u2t would become a really exiting and
>> brilliant tool.
>
> This is an interesting blog from a couple of years ago regarding
> UML2Tools and Papyrus.
>
> http://kenn-hussey.blogspot.com/2008/04/on-doing-right-thing .html
>
> The most interesting bit is the discussion where, effectively, my
> interpretation is that they're saying not only what we're basically
> talking about here (effectively replication of effort) but that there
> seems to be a view that Papyrus is using some of the work from
> UML2Tools but that UML2Tools was still in incubation then, and it
> seemed odd to allow a new project to build on something that's still
> in incubation!
>
>> But as long as everbody thinks to be so much smarter we will repeat
>> all mistakes and errors again and again and won't be one step closer
>> to a good and feature-complete UML2 Tool for Eclipse.
>
> Yes. The blog post above has a comment in it somewhere that suggest
> that perhaps it would have been better for the UML2Tools people and
> the Papyrus people to work together as one (at least that's my
> impression of it).
>
|
|
| |
| Re: I'm worried - UML2Tools vs Papyrus [message #626012 is a reply to message #558506] |
Mon, 13 September 2010 14:36 |
Ed Merks
Messages: 33137
Registered: July 2009 |
Senior Member |
|
|
John,
Comments below.
John McCabe wrote:
> As I understand it, UML2Tools and Papyrus, in principle seem to do
> much the same job. Some time ago I found a UML tool plugin for Eclipse
> that got to a certain point then went commercial. It was a long time
> later that I started looking again and found UML2Tools which seemed to
> fit the bill. However, as others have noted, it seems to have stalled
> without providing an easy route to install in Helios. Papyrus on the
> other hand appears to be actively developed but is not as advanced as
> UML2Tools.
>
> This, to me, is the crux of one of the main issues with Open Source
> Development (and the like). Even if someone does make UML2Tools work
> in Helios, it would appear that it can't be guaranteed that any
> further development will take place.
>
The only way to guarantee ongoing open source development is to do it
yourself or to pay someone to do it. The fact that it's open source
enables that possibility.
> To me there seems to be an awful lot of ego involved in software tool
> development these days.
Egos are involved in all aspects of human social lives.
> Too many people thinking they can do something
> better than the last person who tried and go off on their own merry
> way rather than trying to get involved with the last person who tried
> and helping to shape the product more to their needs (or the needs of
> the people they've surveyed about the product).
>
Of course. Not only that, each generation seems determined to repeat the
mistakes of the previous one.
> It seems to me that the only way round this is commercial product
> development!
>
No, it's fundamentally the same ego problem there. Of course the
difference there is that money exchanges hands so you can demand that
you get value for it. Choose just as carefully though. Many commercial
products also go the way of the dodo, and those that don't will be
revved and modernized so you'll be spending your money again and
learning how to use it again...
> In the meantime though; if I wanted to use UML in Eclipse, what is the
> best option?
>
At this point, Papyrus seems the most active and the most likely to
remain that way.
> John
>
Ed Merks
Professional Support: https://www.macromodeling.com/
|
|
|
| Re: I'm worried - UML2Tools vs Papyrus [message #626013 is a reply to message #558578] |
Mon, 13 September 2010 15:51 |
John McCabe
Messages: 228
Registered: July 2009 |
Senior Member |
|
|
On Mon, 13 Sep 2010 07:36:22 -0700, Ed Merks <Ed.Merks@gmail.com>
wrote:
Ed - thanks for your comments. See below.
<..snip..>
>> This, to me, is the crux of one of the main issues with Open Source
>> Development (and the like). Even if someone does make UML2Tools work
>> in Helios, it would appear that it can't be guaranteed that any
>> further development will take place.
>>
>The only way to guarantee ongoing open source development is to do it
>yourself or to pay someone to do it. The fact that it's open source
>enables that possibility.
That's very true. From a commercial viewpoint though (at least, as a
commercial entity trying to use OSS), this requires investment in one
way or another; either you get someone else to do it, in which case
hopefully you're one of many customers so you basically share the cost
of the development, or you do it in house in which case your
cost/benefit analysis shows a potentially less attractive outcome.
>> To me there seems to be an awful lot of ego involved in software tool
>> development these days.
>Egos are involved in all aspects of human social lives.
Well true, but given the emphasis over the 23 years or so I've been a
professional software developer has been strongly on teamwork, it
would have been nice to believe that, somewhere along the line,
someone may have been paying attention :-) It's ironic (to me) that
many OSS projects are aimed at helping teams work more effectively,
yet the developers working on those projects have chosen not to join
an existing team and, as you say, have ended up repeating the mistakes
of previous projects. My own view of UML in Eclipse seems to support
that!
>> Too many people thinking they can do something
>> better than the last person who tried and go off on their own merry
>> way rather than trying to get involved with the last person who tried
>> and helping to shape the product more to their needs (or the needs of
>> the people they've surveyed about the product).
>Of course. Not only that, each generation seems determined to repeat the
>mistakes of the previous one.
>> It seems to me that the only way round this is commercial product
>> development!
>No, it's fundamentally the same ego problem there. Of course the
>difference there is that money exchanges hands so you can demand that
>you get value for it.
I'm not convinced it's fundamentally the same ego problem. From a
commercial point of view you might believe you can do a better job
than a competitor but that wouldn't necessarily mean you would go
ahead and do it. There are so many variables involved; you need to
know there is a commercial market for another product, you'd have to
make sure you could do it within time and budget constraints yet still
be better than the opposition. It's not just a matter of a guy sitting
in his bedroom deciding "I'll do a better one of those" and going off
on one.
FWIW - programming languages are a case in point there; the
proliferation of different languages these days is, to me, ridiculous.
If, rather than just inventing a new language, some of these talented
people just got involved in working on one of the existing languages
to try to improve it in the particular paradigm they're interested in,
the world would be a much better place :-)
>Choose just as carefully though. Many commercial
>products also go the way of the dodo, and those that don't will be
>revved and modernized so you'll be spending your money again and
>learning how to use it again...
This is true of course, but there are still a number of products that
have been developed in to good, stable commercial products but their
'authors' have decided to abandon them. In some cases, e.g.
OpenWatcom, Delphi, etc they seem to have found their way in to the
public domain for future development and still have some support from
the original development team or at least some of the original
development team!
>> In the meantime though; if I wanted to use UML in Eclipse, what is the
>> best option?
>At this point, Papyrus seems the most active and the most likely to
>remain that way.
Thanks for that. If I'd asked 2 years ago I suspect you could have
replaced Papyrus with UML2Tools :-)
All the best
John
|
|
| | | | |
| Re: I'm worried - UML2Tools vs Papyrus [message #626092 is a reply to message #559000] |
Thu, 16 September 2010 20:28 |
|
Hi all,
My two cents on that subject as I have been involved in the creation of
MDT Papyrus.
1. 2007
In 2007, UMLTools was the only UML2 tool hosted by the Eclipse
Foundation and the Eclipse community hoped it was going to become THE
reference open source UML tool.
Many companies started to use it and to extend it to create modeling
platforms.
In the TOPCASED project, we studied the possibility to rely on GMF
runtime and perhaps on UML2 tools to improve TOPCASED modeling editors.
We faced some issues and submited patches. None of them were analysed by
GMF and by UML2Tools teams during 8 months.
2. EclipseCon 2008
At EclipseCon 2008, there were many participants interested by a UML2
graphical editor hosted by the foundation and we were several different
people to complain that there was no real effort from the Borland guys
to take our patches into account.
There was a "Bird Of a feather" session one evening in which the debate
was strong concerning the strategy of the Borland team, who admitted
that they had no time to consider our patches and had no volonty to
improve UML2Tools for two reasons:
a) UML2Tools was only a test case for GMF : prove that you can
generate 100% of UML2 diagrams with GMF
b) BORLAND had a commercial product based on GMF and they did not want
UML2Tools to provide some competitive functionalities.
So it was then clear that UML2tools was not planned to become a tool of
industrial quality.
If some people were ready to put efforts on an other UML2 tool, the
place was left open.
3. September 2008 : the advent of MDT Papyrus
TOPCASED guys, Papyrus 1 guys (CEA) and Moskitt guys (spanish
initiative) were ready to put efforts to create something of industrial
quality on the basis of UML2 tools.
We started the MDT Papyrus component with the goal to reach industrial
quality level as soon as possible.
4. Why not reusing UML2Tools for MDT Papyrus?
We started reusing UML2Tools. But it was very very hard to maintain as
tha main logics were embedded into the GMFGen templates. We wanted a
modular approach with a "backbone" of services able to provide
transversal services : model explorer, split models, share diagrams with
a single model... all things that were simply not possible with UML2Tools.
After 5 months we decided not to use UM2Tools as foundation as we lost
too much time trying to reuse it...
Sure, you might consider that time has been lost by doing things twice
but they were done in a completely differnt goal and MDT Papyrus aims at
delivering full support from UML2 and industrial quality.
We did some errors by putting two many things in the 0.7.0 release and
spending enough time on validation and bug fixes and performance but we
are going to fix those issues to reach a good quality level in a few months.
5. conclusion
I'm sorry for people who invested a lot on extending UML2Tools and
discover that this is not really maintained but if some of you are
really active then just give us your feedback concerning MDT Papyrus and
you will be given the opportunity to contribute and perhaps become
committer after a few patches.
Welcome.
Hope this context and history helps in understanding this situation,
kind Regards
Raphaël
TOPCASED and MDT Papyrus team
Le 15/09/2010 11:26, John McCabe a écrit :
> On Wed, 15 Sep 2010 03:45:38 -0400, Christian Waniek
> <chris.waniek@gmx.net> wrote:
>
>> I really aggree about this with you. But nevertheless are we using
>> UML2 Tools in our company. So I'm on it to do some develoment for
>> U2T (hoping to become a comitter soon). Because we want to push this
>> project.
>
> Good on you. I wish you the best of luck.
>
>> Although actually it looks like U2T will die, I'm convinced this won't
>> happen because U2T has high potential and there is so much work
>> already done...and the project lead of U2T tries to push this project.
>
>> So I also can't understand why there has to be a new project which
>> actually has the same goal. If you would have invested in u2t as much
>> as there has been to papyrus, u2t would become a really exiting and
>> brilliant tool.
>
> This is an interesting blog from a couple of years ago regarding
> UML2Tools and Papyrus.
>
> http://kenn-hussey.blogspot.com/2008/04/on-doing-right-thing .html
>
> The most interesting bit is the discussion where, effectively, my
> interpretation is that they're saying not only what we're basically
> talking about here (effectively replication of effort) but that there
> seems to be a view that Papyrus is using some of the work from
> UML2Tools but that UML2Tools was still in incubation then, and it
> seemed odd to allow a new project to build on something that's still
> in incubation!
>
>> But as long as everbody thinks to be so much smarter we will repeat
>> all mistakes and errors again and again and won't be one step closer
>> to a good and feature-complete UML2 Tool for Eclipse.
>
> Yes. The blog post above has a comment in it somewhere that suggest
> that perhaps it would have been better for the UML2Tools people and
> the Papyrus people to work together as one (at least that's my
> impression of it).
>
|
|
| |
| Re: I'm worried - UML2Tools vs Papyrus [message #628044 is a reply to message #627968] |
Thu, 23 September 2010 08:46 |
Ed Willink
Messages: 7655
Registered: July 2009 |
Senior Member |
|
|
Hi Christian
> And how was it possible that Borland blocked the advancement of U2T?
> (because it is open source) Why did nobody escalate this problem.Because
> there isn't only a project lead of U2T, there is one at MDT level and
> there is one for the whole Modeling-area in Eclipse.
Without referring to this particular case, there is a big problem but it
is perhaps unavoidable.
Projects need committers most of whom need financial income, so
committers tend to work for companies who are interested in the success
of their work.
Projects need coherent structure so new committers cannot be added
quickly and arbitrarily.
When projects have committers from a very small number, perhaps only
one, of companies a potential for conflict arises between the company
and community interests.
There are a number of very successful one company projects, so we have
to live with the risk. Unfortunately when a project doesn't evolve in
the direction that someone wants, it takes a long time to determine that
a direction is being unreasonably thwarted and even longer to
demonstrate that agreed remedial responses are not being adhered to. So
if you want your direction to proceed today rather than next year, you
are left with no choice but to compete.
I suspect that solutions to this problem may be worse than the problem.
Eclipse is Open so alternative projects can compete and do in a number
of areas.
UML2-Tools is an incubation project. This means that it has not achieved
sufficient standaing to be relied on. Users do so at their own risk.
The only solution I see would be for committers to be employed by
Eclipse, which would then require a substantial levy on at least
companies and possibly users. There would be a massive planning
bureaucracy that would make decisions that few people would agree with.
Eclipse would no longer be free. Not really an option.
Regards
Ed Willink
|
|
| |
Goto Forum:
Current Time: Fri Apr 19 08:35:46 GMT 2024
Powered by FUDForum. Page generated in 0.02193 seconds |
] 
Search
Help
Register
Login
Home
