[openstack-dev] MuranoPL questions?

Joshua Harlow harlowja at yahoo-inc.com
Tue Mar 11 22:05:57 UTC 2014


I guess I might be a bit biased to programming; so maybe I'm not the target audience.

I'm not exactly against DSL's, I just think that DSL's need to be really really proven to become useful (in general this applies to any language that 'joe' comp-sci student can create). Its not that hard to just make one, but the real hard part is making one that people actually like and use and survives the test of time. That’s why I think its just nicer to use languages that have stood the test of time already (if we can), creating a new DSL (muranoPL seems to be slightly more than a DSL imho) means creating a new language that has not stood the test of time (in terms of lifetime, battle tested, supported over years) so that’s just the concern I have.

Of course we have to accept innovation and I hope that the DSL/s makes it easier/simpler, I just tend to be a bit more pragmatic maybe in this area.

Here's hoping for the best! :-)

-Josh

From: Renat Akhmerov <rakhmerov at mirantis.com<mailto:rakhmerov at mirantis.com>>
Reply-To: "OpenStack Development Mailing List (not for usage questions)" <openstack-dev at lists.openstack.org<mailto:openstack-dev at lists.openstack.org>>
Date: Monday, March 10, 2014 at 8:36 PM
To: "OpenStack Development Mailing List (not for usage questions)" <openstack-dev at lists.openstack.org<mailto:openstack-dev at lists.openstack.org>>
Subject: Re: [openstack-dev] MuranoPL questions?

Although being a little bit verbose it makes a lot of sense to me.

@Joshua,

Even assuming Python could be sandboxed and whatever else that’s needed to be able to use it as DSL (for something like Mistral, Murano or Heat) is done  why do you think Python would be a better alternative for people who don’t know neither these new DSLs nor Python itself. Especially, given the fact that Python has A LOT of things that they’d never use. I know many people who have been programming in Python for a while and they admit they don’t know all the nuances of Python and actually use 30-40% of all of its capabilities. Even not in domain specific development. So narrowing a feature set that a language provides and limiting it to a certain domain vocabulary is what helps people solve tasks of that specific domain much easier and in the most expressive natural way. Without having to learn tons and tons of details that a general purpose language (GPL, hah :) ) provides (btw, the reason to write thick books).

I agree with Stan, if you begin to use a technology you’ll have to learn something anyway, be it TaskFlow API and principles or DSL. Well-designed DSL just encapsulates essential principles of a system it is used for. By learning DSL you’re leaning the system itself, as simple as that.

Renat Akhmerov
@ Mirantis Inc.



On 10 Mar 2014, at 05:35, Stan Lagun <slagun at mirantis.com<mailto:slagun at mirantis.com>> wrote:

> I'd be very interested in knowing the resource controls u plan to add. Memory, CPU...
We haven't discussed it yet. Any suggestions are welcomed

> I'm still trying to figure out where something like https://github.com/istalker2/MuranoDsl/blob/master/meta/com.mirantis.murano.demoApp.DemoInstance/manifest.yaml would be beneficial, why not > just spend effort sand boxing lua, python... Instead of spending effort on creating a new language and then having to sandbox it as well... Especially if u picked languages that are made to be  sandboxed from the start (not python)...

1. See my detailed answer in Mistral thread why haven't we used any of those languages. There are many reasons besides sandboxing.

2. You don't need to sandbox MuranoPL. Sandboxing is restricting some operations. In MuranoPL ALL operations (including operators in expressions, functions, methods etc.) are just those that you explicitly provided. So there is nothing to restrict. There are no builtins that throw AccessViolationError

3. Most of the value of MuranoPL comes not form the workflow code but from class declarations. In all OOP languages classes are just a convenient to organize your code. There are classes that represent real-life objects and classes that are nothing more than data-structures, DTOs etc. In Murano classes in MuranoPL are deployable entities like Heat resources application components, services etc. In dashboard UI user works with those entities. He (in UI!) creates instances of those classes, fills their property values, binds objects together (assigns on object to a property of another). And this is done without even a single MuranoPL line being executed! That is possible because everything in MuranoPL is a subject to declaration and because it is just plain YAML anyone can easily extract those declarations from MuranoPL classes.
Now suppose it was Python instead of MuranoPL. Then you would have to parse *.py files to get list of declared classes (without executing anything). Suppose that you managed to solve this somehow. Probably you wrote regexp that finds all class declarations it text files. Are you done? No! There are no properties (attributes) declarations in Python. You cannot infer all possible class attributes just from class declaration. You can try to parse __init__ method to find attribute initialization code but then you find that you cannot infer property types. You would not know what values does class expect in those attributes, what are constraints etc. Besides there are plenty of ways to fool such naive algorithms. Classes can be created using type() built-in function without declarations at all. Attributes may be initialized anywhere. Everything can be made in runtime.
As you can see Python is not good for the task. Ans all this true for all of the languages you mentioned. The reason is fundamental - all of those languages are dynamic.
So maybe you can take some statically-typed language like C++ or Java to do the job? Again, no! There are other problems with those type of languages. Have you ever seen statically-typed embeddable language? Those language require compilation and linker, binary distribution of classes, handling of versioning problems, dependencies on 3-rd party libraries etc. This is the hell you don't want to step in. But even in statically typed languages there are no contracts. You know the list of properties and their types. But knowing that class foo has property bar of type int doesn't give a clue if value if 10 is okay for that property.

Now I admit that we could probably spend a year of development and try to implement this on type Java with custom wrapping and heavy usage of decorators, embed somehow JVM into python process, do all possible sandboxing and then get a Frankenstein that nobody in OpenStack community would like or accept.

My point here is that there is no language in the world that was designed for external object composition that is in Murano. No language in the world was design so that there would be huge repository of classes but no single control point, no main project, no trust between different classes. There is no language with tree-like heap organization (each object has single owner object). This what makes MuranoPL domain-specific and not deployments. There are no such languages not because nobody that smart to create them but because nobody needed to do so. Because MuranoPL is domain-specific unlike Python or Java.

MuranoPL does contain many constructs that can be found in general-purpose languages but that doesn't make it general-purpose!


> Who is going to train people on muranoPL, write language books and tutorials when the same amount of work has already been done for 10+ years for other languages....
Let me say that most of MuranoPL audience are bash-developers that know Lua to the same level they know MuranoPL :)

In Mirantis we deed interesting experiment. We asked several deployment engineers that were not involved in MuranoPL creation and knew nothing about it beforehand to write some application manifests using MuranoPL. We gave them 10-minute introduction, link to DSL documentation blueprint and examples in my github repository (you've seen both of them). The link above to DemoInstance manifest points to the class that was created by one of those guys. It took them several days to get used to it. But they never even used Python before! And the feedback was very positive.

I don't really see a problem writing tutorials or even a book. Learning new framework like TaskFlow is nowhere easier than learning DSL. Anyway you need to learn something to start using new technology. There are hundreds of DSL languages that are used in production and I don't feel like having one more language is really a problem




On Mon, Mar 10, 2014 at 12:26 AM, Joshua Harlow <harlowja at yahoo-inc.com<mailto:harlowja at yahoo-inc.com>> wrote:
I'd be very interested in knowing the resource controls u plan to add. Memory, CPU...

I'm still trying to figure out where something like https://github.com/istalker2/MuranoDsl/blob/master/meta/com.mirantis.murano.demoApp.DemoInstance/manifest.yaml would be beneficial, why not just spend effort sand boxing lua, python... Instead of spending effort on creating a new language and then having to sandbox it as well... Especially if u picked languages that are made to be sandboxed from the start (not python)...

Who is going to train people on muranoPL, write language books and tutorials when the same amount of work has already been done for 10+ years for other languages....

I fail to see where muranoPL is a DSL when it contains a full language already with functions, objects, namespaces, conditionals... (what is the domain of it?), maybe I'm just confused though (quite possible, haha).

Does this not seem awkward to anyone else??

Sent from my really tiny device...

On Mar 8, 2014, at 10:44 PM, "Stan Lagun" <slagun at mirantis.com<mailto:slagun at mirantis.com>> wrote:

First of all MuranoPL is not a host but hosted language. It never intended to replace Python and if Python can do the job it is probably better than MuranoPL for that job.
The problem with Python is that you cannot have Python code as a part of your DSL if you need to evaluate that DSL on server-side. Using Python's eval() is not secure. And you don't have enough control on what evaled code is allowed to do. MuranoPL on the contrary are fully sandboxed. You have absolute control over what functions/methods/APIs are exposed to DSL and DSL code can do nothing except for what it allowed to do. Besides you typically do want your DSL to be domain-specific so general-purpose language like Python can be suboptimal.

I don't say MuranoPL is good for all projects. It has many Murano-specific things after all. In most cases you don't need all those OOP features that MuranoPL has. But the code organization, how it uses YAML, block structures and especially YAQL expressions can be of a great value to many projects

For examples of MuranoPL classes you can browse https://github.com/istalker2/MuranoDsl/tree/master/meta folder. This is my private repository that I was using to develop PoC for MuranoPL engine. We are on the way to create production-quality implementation with unit-tests etc. in regular Murano repositories on stackforge


On Sun, Mar 9, 2014 at 7:33 AM, Joshua Harlow <harlowja at yahoo-inc.com<mailto:harlowja at yahoo-inc.com>> wrote:
To continue from other thread....

"""
Personally I believe that YAQL-based DSLs similar (but probably simlet than) MuranoPL can be of a great value for many OpenStack projects that have DSLs of different kind. Murano for App Catatalog, Mistral for workflows, Heat for HOT templates, Ceilometer for alarms & counter aggregation rules, Nova for customizable resource scheduling and probably many more.
"""

Isn't python a better host language for said DSLs than muranoPL? I am still pretty weary of a DSL that starts to grow so many features to encompass other DSLs since then it's not really a DSL but a non-DSL, and we already have one that everyone is familiar with (python).

Are there any good examples if muranoPL that I can look at that take advantage of muranoPL classes, functions, namespaces which can be compared to there python equivalents. Has such an analysis/evaluation been done?

Sent from my really tiny device...
_______________________________________________
OpenStack-dev mailing list
OpenStack-dev at lists.openstack.org<mailto:OpenStack-dev at lists.openstack.org>
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev



--
Sincerely yours
Stanislav (Stan) Lagun
Senior Developer
Mirantis
35b/3, Vorontsovskaya St.
Moscow, Russia
Skype: stanlagun
www.mirantis.com<http://www.mirantis.com/>
slagun at mirantis.com<mailto:slagun at mirantis.com>
_______________________________________________
OpenStack-dev mailing list
OpenStack-dev at lists.openstack.org<mailto:OpenStack-dev at lists.openstack.org>
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

_______________________________________________
OpenStack-dev mailing list
OpenStack-dev at lists.openstack.org<mailto:OpenStack-dev at lists.openstack.org>
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev




--
Sincerely yours
Stanislav (Stan) Lagun
Senior Developer
Mirantis
35b/3, Vorontsovskaya St.
Moscow, Russia
Skype: stanlagun
www.mirantis.com<http://www.mirantis.com/>
slagun at mirantis.com<mailto:slagun at mirantis.com>
_______________________________________________
OpenStack-dev mailing list
OpenStack-dev at lists.openstack.org<mailto:OpenStack-dev at lists.openstack.org>
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20140311/d5e4ca3d/attachment.html>


More information about the OpenStack-dev mailing list