aHuman HLAv3
41 views
Skip to first unread message
Habbit
Mar 22, 2011, 4:41:48 AM3/22/11
to Discuss a Human Project
Hi all!
Let me congratulate us with new stage in development - HLAv3.
I'm sure it opens the way to project goals and allows to apply efforts
more steadily.
What differs in HLAv3 from HLAv2:
1. It is the first time when HLA was named.
It makes easier to speak about HLA decisions and define further
project design.
It gives way to speak about intentional changes we do, compare current
architecture with previous one.
I allows to speak about switching to next version of architecture when
we feel it is required and this it just part of internal process, not
unexpected mess, killing the project development
2. HLAv3 introduces actuators concept
It was in my mind and known before, but HLAv2 required sensors and
effectors combined with body, which prevents from seeing its
relationship with executable targets.
Also chat study showed a problem of building generic approach for
sensors and effectors.
Chat session is the only channel whilst one direction is sensing and
another direction is executing action.
Now specific actuator set can be implemented adjusted to specific
project target without trying to be generic
3. HLAv3 makes difference between ready-to-use libraries, universal
programming techniques, mind techniques and final project targets.
Previously we had actuators mixed with mind and it makes a mess when
we do not understand what executable should do.
It makes 2-level structure of HLA - first states that we have
libraries, platform, mind and actuator sets architectural building
blocks. Second defines HLA for each building block.
Btw, HLA for libraries and actuators are not fully defined there.
4. HLAv3 introduces layered mind model
After detailed studying of real human mind I understood its multi-
layered construction (area, region and neural layers) and existence of
several domains of execution - e.g. cortex and nucleous.
These layers can be undergo research, coded and packaged separately.
It has led me to defining mind layers and several types of modules
related to the layer.
In the same time whole aHuman architecture is not layered.
5. HLAv3 introduces component architecture of aHuman platform
Previously it was a mix of helper classes.
Component model will simplify understanding the platform and its usage
and tracking.
6. Last but not least, level of understanding biological being is now
comparable to level of understanding AI techniques.
With this, and I hope, team will join me in such efforts, project will
be on track of feasible and plausible implementation.
We can do some things the same way as biological mind, some -
differently, but now we can speak about intentional changes in fact-
based style.
What is definitely imperfect in current architecture page:
- limitations section is copied from HLAv2 and by chance includes non-
justified design decisions
- platform HLA has been created in haste - splitting needs to be
better
- library building block has no HLA
- no picture of mind in this architectural approach
Let me congratulate us with new stage in development - HLAv3.
I'm sure it opens the way to project goals and allows to apply efforts
more steadily.
What differs in HLAv3 from HLAv2:
1. It is the first time when HLA was named.
It makes easier to speak about HLA decisions and define further
project design.
It gives way to speak about intentional changes we do, compare current
architecture with previous one.
I allows to speak about switching to next version of architecture when
we feel it is required and this it just part of internal process, not
unexpected mess, killing the project development
2. HLAv3 introduces actuators concept
It was in my mind and known before, but HLAv2 required sensors and
effectors combined with body, which prevents from seeing its
relationship with executable targets.
Also chat study showed a problem of building generic approach for
sensors and effectors.
Chat session is the only channel whilst one direction is sensing and
another direction is executing action.
Now specific actuator set can be implemented adjusted to specific
project target without trying to be generic
3. HLAv3 makes difference between ready-to-use libraries, universal
programming techniques, mind techniques and final project targets.
Previously we had actuators mixed with mind and it makes a mess when
we do not understand what executable should do.
It makes 2-level structure of HLA - first states that we have
libraries, platform, mind and actuator sets architectural building
blocks. Second defines HLA for each building block.
Btw, HLA for libraries and actuators are not fully defined there.
4. HLAv3 introduces layered mind model
After detailed studying of real human mind I understood its multi-
layered construction (area, region and neural layers) and existence of
several domains of execution - e.g. cortex and nucleous.
These layers can be undergo research, coded and packaged separately.
It has led me to defining mind layers and several types of modules
related to the layer.
In the same time whole aHuman architecture is not layered.
5. HLAv3 introduces component architecture of aHuman platform
Previously it was a mix of helper classes.
Component model will simplify understanding the platform and its usage
and tracking.
6. Last but not least, level of understanding biological being is now
comparable to level of understanding AI techniques.
With this, and I hope, team will join me in such efforts, project will
be on track of feasible and plausible implementation.
We can do some things the same way as biological mind, some -
differently, but now we can speak about intentional changes in fact-
based style.
What is definitely imperfect in current architecture page:
- limitations section is copied from HLAv2 and by chance includes non-
justified design decisions
- platform HLA has been created in haste - splitting needs to be
better
- library building block has no HLA
- no picture of mind in this architectural approach
Habbit
Mar 22, 2011, 4:57:34 AM3/22/11
to Discuss a Human Project
Please review architecture this week:
- change contents of architecture page without asking if you are sure
(to improve wording, add concepts, correct concepts) - I will review
updates and contact you if required
- ask questions here - I will answer
- propose here some changes if you are not sure or do not know how
better to add it to the page
- change contents of architecture page without asking if you are sure
(to improve wording, add concepts, correct concepts) - I will review
updates and contact you if required
- ask questions here - I will answer
- propose here some changes if you are not sure or do not know how
better to add it to the page
Phoenix
Mar 22, 2011, 8:59:53 AM3/22/11
to Discuss a Human Project
Hi Habbit,
I think this architecture is more refined now and clearly defines
which module should do what instead of what we were doing till now. We
might need to define the architecture for brain separately as it is
the most complex system in this project. May be the block diagrams or
flow charts as to how the information will flow.
now we are 3 in numbers so I will recommend using the software
development approach where we have a rough SRS document and each
developer works on his own module and implement the functionality of
the module depending on the SRS information. I am not suggesting that
we should formally create a SRS but a rough guideline will be enough.
Let me know your views about this.
Thank you for all the efforts you put in this project, I appreciate
your efforts.
Thanks,
Sarbjit
I think this architecture is more refined now and clearly defines
which module should do what instead of what we were doing till now. We
might need to define the architecture for brain separately as it is
the most complex system in this project. May be the block diagrams or
flow charts as to how the information will flow.
now we are 3 in numbers so I will recommend using the software
development approach where we have a rough SRS document and each
developer works on his own module and implement the functionality of
the module depending on the SRS information. I am not suggesting that
we should formally create a SRS but a rough guideline will be enough.
Let me know your views about this.
Thank you for all the efforts you put in this project, I appreciate
your efforts.
Thanks,
Sarbjit
quentin valdeyron
Mar 22, 2011, 1:22:09 PM3/22/11
to ahuman-...@googlegroups.com
Hi everyone,
I still have problems to understand some parts of the code, only because I don't measure how usefull they can be and why they are that way.
That's why I have begun to work on UML diagram of ahuman' architecture concepts.
Cheers
Quentin,
--
You received this message because you are subscribed to the Google
Groups "Discuss a Human Project" group.
To post to this group, send email to ahuman-...@googlegroups.com
Habbit
Mar 22, 2011, 5:25:50 PM3/22/11
to Discuss a Human Project
Need one more day to finish refactoring.
Too many items to change.
Too many items to change.
Habbit
Mar 22, 2011, 5:27:22 PM3/22/11
to Discuss a Human Project
Not sure, that my post was added - as I see it in my email, not not
here :(
Hence maybe duplicate one:
I agree about idea of SRS & UML, providing that:
1. Component diagram for HLA is quite desired, whilst sequence/flow
diagram can be not feasible, because of:
- biological model is not fully completed now and contains
contradictory information (it should be refined with time)
- HLA introduces types of mind areas/regions, but not specific list of
them
btw, what tool will you use for UML diagram?
2. SRS is created by developer in parallel with design and/or coding:
- SRS is a kind of technical requirements, usually a product of
elaboration stage, and fully completed only by the alpha readiness
- SRS differs from business requirements, which are created exactly
before design/coding efforts to start
- business requirements for tasks with significant research efforts
are set by the end of proof-of-the-concept/prototyping stage
Therefore, we can speak about:
a) setting preliminary business requirements - exactly before doing
any research/design (I can do it now)
b) creating business requirements within certain period of time when
most active research occurs (I can do it when major design options
will be known)
c) creating technical requirements, better be done by streamline owner
in parallel with coding - and review from other team members
Regards,
Vladimir
here :(
Hence maybe duplicate one:
I agree about idea of SRS & UML, providing that:
1. Component diagram for HLA is quite desired, whilst sequence/flow
diagram can be not feasible, because of:
- biological model is not fully completed now and contains
contradictory information (it should be refined with time)
- HLA introduces types of mind areas/regions, but not specific list of
them
btw, what tool will you use for UML diagram?
2. SRS is created by developer in parallel with design and/or coding:
- SRS is a kind of technical requirements, usually a product of
elaboration stage, and fully completed only by the alpha readiness
- SRS differs from business requirements, which are created exactly
before design/coding efforts to start
- business requirements for tasks with significant research efforts
are set by the end of proof-of-the-concept/prototyping stage
Therefore, we can speak about:
a) setting preliminary business requirements - exactly before doing
any research/design (I can do it now)
b) creating business requirements within certain period of time when
most active research occurs (I can do it when major design options
will be known)
c) creating technical requirements, better be done by streamline owner
in parallel with coding - and review from other team members
Regards,
Vladimir
Habbit
Mar 22, 2011, 5:19:43 PM3/22/11
to Discuss a Human Project
Habbit
Mar 23, 2011, 3:46:37 PM3/23/11
to Discuss a Human Project
Commited first set of files in HLAv3:
1. All architectural items are exposed, except mind HLA.
2. Created externals solution workspace with generic library
3. Created aHuman HLAv3 solution worspace with tools, library,
platform, mind and aWee projects
4. Library project contains binary version of generic library from
externals build
5. Tools project contains folder for aiconsole - TBD
6. Platform project contains all components, but not migrated
messaging, media and db (last maybe ignored), also does not contains
configuring - which TND in EnvService.
7. Mind project is only folder, but with options to generate library
in proper place
8. aWee project is executable for aWee target. It is able to build and
run.
Now all commited is clean code, works as desired - exception handling
works, stack trace works, simple logging works.
I will continue to add remaining, but project is already available for
review & UML.
Going to finish migration tomorrow.
1. All architectural items are exposed, except mind HLA.
2. Created externals solution workspace with generic library
3. Created aHuman HLAv3 solution worspace with tools, library,
platform, mind and aWee projects
4. Library project contains binary version of generic library from
externals build
5. Tools project contains folder for aiconsole - TBD
6. Platform project contains all components, but not migrated
messaging, media and db (last maybe ignored), also does not contains
configuring - which TND in EnvService.
7. Mind project is only folder, but with options to generate library
in proper place
8. aWee project is executable for aWee target. It is able to build and
run.
Now all commited is clean code, works as desired - exception handling
works, stack trace works, simple logging works.
I will continue to add remaining, but project is already available for
review & UML.
Going to finish migration tomorrow.
Reply all
Reply to author
Forward
0 new messages