Graph vendor Implementation

352 views
Skip to first unread message

Janardhan Reddy

unread,
Oct 31, 2016, 7:16:20 AM10/31/16
to Gremlin-users
Hi,

What is the best way to integrate tinkerpop into a go based graph vendor(Dgraph). 

1. One of the ways is to implement tinkerpop core api, but then the clients can use the graph only in embedded mode. Else we need to run gremlin server on top of graph database to which the clients can connect to.

2. The documentation for RemoteConnection enabled graph system is not clear. How does it work.

Is it possible to enable tinkerpop enabled stacks(graph language providers etc) to pass gremlin bytecode to remote connection enabled graph system and the graph vendor would take care of parsing and executing the query.

Marko Rodriguez

unread,
Oct 31, 2016, 12:14:49 PM10/31/16
to gremli...@googlegroups.com
Hello,

> What is the best way to integrate tinkerpop into a go based graph vendor(Dgraph).

This is a hard problem.

Gremlin is both a virtual machine and a language.

The only (known) implementation of the Gremlin virtual machine is in Java for the JVM. This is the Gremlin machine distributed by Apache TinkerPop. Someone can implement the Gremlin machine in Go, but it would be a bit of an undertaking. However, for someone to build a Gremlin language variant in Go, its quite easy. They just need to generate Gremlin bytecode and be able to serialize it (via GraphSON — json) over the wire to the Gremlin machine.

> Is it possible to enable tinkerpop enabled stacks(graph language providers etc) to pass gremlin bytecode to remote connection enabled graph system and the graph vendor would take care of parsing and executing the query.

Yes, but the problem is that you will have to know how to process Gremlin bytecode which is basically building a Gremlin machine in Go. Its not hard, its just lots of code. What would be best is if you could translate Gremlin bytecode to DGraph’s query language. Then your architecture would be:

Gremlin language variant —> Gremlin bytecode —> RemoteConnection —> DGraph query language —> DGraph

In this way, ANY Gremlin language variant could take to DGraph. The problem, this solution is tied to DGraph and not generalized to work for a Go-based graph database.

I can answer more questions if you have them…

Good luck,
Marko.

Janardhan Reddy

unread,
Oct 31, 2016, 2:15:34 PM10/31/16
to Gremlin-users
Hi,

I was actually looking a solution for dgraph only and was thinking of  converting gremlin bytecode to dgraph's subgraph so that dgraph can execute it.

Can you please point me to the documentation of gremlin bytecode and remoteconnection. How does any gremlin language variant talk via remote connection( I was going through the gremlin server code and it seemed like the remote connection uses http).

Janardhan Reddy

unread,
Oct 31, 2016, 2:20:36 PM10/31/16
to Gremlin-users
Can you please share your thoughts on integrating gremlin(tinkerpop) in dgraph so that graph traversal control is in dgraph rather than through tinkerpop api.


On Monday, 31 October 2016 21:44:49 UTC+5:30, Marko A. Rodriguez wrote:

Janardhan Reddy

unread,
Oct 31, 2016, 2:32:01 PM10/31/16
to Gremlin-users
Does the language variants optimise gremlin byte code and pass to remote connection ? What is the best way to take advantage of all tinkerpop features while having control over traversal.

Marko Rodriguez

unread,
Nov 1, 2016, 8:47:13 AM11/1/16
to gremli...@googlegroups.com
Hello,

Does the language variants optimise gremlin byte code and pass to remote connection ?

Currently, no. It is up to the Gremlin traversal machine to do all optimizations. However, there is a ticket for Bytecode-level optimizations.

What is the best way to take advantage of all tinkerpop features while having control over traversal.

The thing is, if the graph database is in Go, you will want a Go-based Gremlin traversal machine. The other option is to say, “I will wrap the DGraph in a Java API that implements TinkerPop’s structure API (Vertex, Edge, etc.).” If you do the latter, you have the problem of Java<->Go bridging, but the benefit of not having to implement a Go-based Gremlin machine.

Keep the questions coming if you have them.

Marko.



On Monday, 31 October 2016 16:46:20 UTC+5:30, Janardhan Reddy wrote:
Hi,

What is the best way to integrate tinkerpop into a go based graph vendor(Dgraph). 

1. One of the ways is to implement tinkerpop core api, but then the clients can use the graph only in embedded mode. Else we need to run gremlin server on top of graph database to which the clients can connect to.

2. The documentation for RemoteConnection enabled graph system is not clear. How does it work.

Is it possible to enable tinkerpop enabled stacks(graph language providers etc) to pass gremlin bytecode to remote connection enabled graph system and the graph vendor would take care of parsing and executing the query.

-- 
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/a81ee7da-f719-473a-b85c-4e495ad7a5c4%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Janardhan Reddy

unread,
Nov 1, 2016, 12:33:40 PM11/1/16
to Gremlin-users
Can you please point me to the documentation of gremlin bytecode and remoteconnection. How does any gremlin language variant talk via remote connection( I was going through the gremlin server code and it seemed like the remote connection uses http).

Robert Dale

unread,
Nov 1, 2016, 1:09:37 PM11/1/16
to Gremlin-users
The gremlin server uses websockets.

For reference, if you decide to implement a graph provider over the dgraph java driver, it would be very similar to what sqlg and neo4j-gremlin-bolt graph providers do.  Their implementations translate into native queries to a remote database: Sqlg -> sql, Neo4j -> cypher. You can find links on the main tinkerpop page.

Janardhan Reddy

unread,
Nov 1, 2016, 1:34:21 PM11/1/16
to Gremlin-users
I had seen neo4j-gremlin-bolt and it seems like only api implementations are translated into native queries to a remote database, i mean the whole query is not translated. So if there are a sequence of traversals won't there be multiple network calls. We would want to avoid that.

pieter-gmail

unread,
Nov 1, 2016, 2:21:11 PM11/1/16
to gremli...@googlegroups.com
Hi,

I can talk a bit about what Sqlg does. Firstly its a java application.
Its best usage is for other java applications.
It translates gremlin/steps to sql. The multiple network calls you
mention is in some ways the primary task of Sqlg. Sqlg collapses
gremlin/steps into as few as possible sql calls.

So g.V().hasLabel("Person").out().out().out() translates to one sql query.

Without this kind of optimization it would be simply be too slow to be
usable.

If you want to build a native Go client then you are in trouble. It is a
huge job to implement all of TinkerPop in another language.

The easy option is to do it in java. Then you build a java client that
optimizes and translates gremlin steps into a few as possible GraphQL
queries. This way you are up and running quickly and you get the benefit
of a number of years of development on the TinkerPop core.

Out of curiosity, does GraphQL and DGraph support properties on edges?
Predicates with properties or references with properties.

Cheers
Pieter
>> <https://groups.google.com/d/msgid/gremlin-users/a81ee7da-f719-473a-b85c-4e495ad7a5c4%40googlegroups.com?utm_medium=email&utm_source=footer>.
>> For more options,
>> visit https://groups.google.com/d/optout
>> <https://groups.google.com/d/optout>.
>
> --
> You received this message because you are subscribed to the Google
> Groups "Gremlin-users" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to gremlin-user...@googlegroups.com
> <mailto:gremlin-user...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/gremlin-users/3c2fcba2-04b5-4e71-b9bd-778f0f8841c5%40googlegroups.com
> <https://groups.google.com/d/msgid/gremlin-users/3c2fcba2-04b5-4e71-b9bd-778f0f8841c5%40googlegroups.com?utm_medium=email&utm_source=footer>.

Marko Rodriguez

unread,
Nov 1, 2016, 9:55:18 PM11/1/16
to gremli...@googlegroups.com
Hi,

I had a thought. 

Let us help you to develop Gremlin(mini) in Go. It is rather trivial to implement a Gremlin machine if you can ignore:

1. SideEffects (global side-effects)
2. Sacks (local side-effects)
3. OLAP
4. Pattern Matching (match())

Basically you need:

1. Traversal
2. Step
3. Traverser
4. Strategy
5. Translator (bytecode -> traversal)
6. Graph API (Vertex, Edge, etc.)
7. I/O (basically GraphSON serializers/deserializers).
— skip OLAP initially or if DGraph's OLAP is Spark/Giraph/Hadoop-etc., then you just route such bytecode submissions to Apache’s Gremlin-Java traversal machine.

The tedium would come in creating all the step implementations: ChooseStep, LabelStep, IdStep, VertexStep, GraphStep, PathStep, etc. etc. What is nice, is you can start small and slowly grow your library of steps… However, once you have those in place, then your traversal machine will be able to accept Gremlin bytecode (from any Gremlin language variant). Given that you would implement Gremlin(mini), certain bytecode instructions would yield an “UnsupportOperationException.” 

Here is a paper the describes the basic architecture and design philosophy behind the Gremlin machine:
Here are some related blog posts:
https://markorodriguez.com/2011/08/03/on-the-nature-of-pipes/ (very old, but nice explanatory graphics)
If you are interested in such a path, I would be willing to write a Gremlin(mini-mini) in Python (lets say) and write a blog post about it so people can learn how to get started.

Finally, if you do go down this route, I strongly advise to use the same terminology and structure as Apache Tinkerpop’s Gremlin machine. With a direct mapping to Go, discussions, advances, etc. will be so much easier…

HTH,
Marko.

Janardhan Reddy

unread,
Nov 2, 2016, 12:56:43 AM11/2/16
to Gremlin-users
Hi,

Thanks for the suggestions. 

I have gone through sqlg and Sqlg collapses 
gremlin/steps into as few as possible sql calls but neo4j-gremlin-bolt doesn't do that. I was just wondering
won't it be too slow to use ?

I am summarising all the options we have to enable tinkerpop support in dgraph.

1.  Build a java client that 
optimizes and translates gremlin steps into Graphql and pass to dgraph. Here we can take 
advantage of gremlin's parser and just need to implement traversal strategy.

Internally Dgraph converts Graphql to a format called subgraph and processes the query.

2. Implement RemoteConnection interface in dgraph which accepts gremlin bytecode via GraphSON(json) 
and let dgraph internally convert it to subgraph and process the query.  We need to take care of parsing and 
should be tedious compared to option 1.

3. Wrapping the DGraph in a Java API that implements TinkerPop’s structure (there would be an issue of java-go binding though) 
and we run a gremlin server on the host for remote connections. But traversal here would be controlled by gremlin through java api's.

4. Implementing gremlin machine in go in dgraph. 

Ideally it would be best if dgraph controls the query execution,
rather than gremlin controlling it through the java api's. 

Example: Say we need to find all likes and dislikes of a person, dgraph shards the graph by predicates rather then vertices, so likes
and dislikes of a person can be found in parallel. It would be tedious to implement all these vendor strategies in gremlin(option 3 and 4)
(It's a duplication of code) and i am not sure whether all optimisations could be implemented in gremlin.

Can you please share your thoughts on pros/cons of each approach. 

Marko Rodriguez

unread,
Nov 2, 2016, 7:58:04 AM11/2/16
to gremli...@googlegroups.com
Hello,

1.  Build a java client that 
optimizes and translates gremlin steps into Graphql and pass to dgraph. Here we can take 
advantage of gremlin's parser and just need to implement traversal strategy.
Internally Dgraph converts Graphql to a format called subgraph and processes the query.

This is your easiest route…and given Pieter’s work on Sqlg to create strategies with “mega steps” (that encapsulate a lot of behavior), you can go far with this.

2. Implement RemoteConnection interface in dgraph which accepts gremlin bytecode via GraphSON(json) 
and let dgraph internally convert it to subgraph and process the query.  We need to take care of parsing and 
should be tedious compared to option 1.

Sounds like a pain in the ass. Basically you are building a “pseudo”-machine. If you do (2), I would just do (4) instead.

3. Wrapping the DGraph in a Java API that implements TinkerPop’s structure (there would be an issue of java-go binding though) 
and we run a gremlin server on the host for remote connections. But traversal here would be controlled by gremlin through java api’s.

This is related to (1) where you would immediately be TinkerPop-compliant and over time you would create more/better strategies to do bulk operations and limit network bandwidth. If you go with (1), I would go with (3).

4. Implementing gremlin machine in go in dgraph. 

Ideally it would be best if dgraph controls the query execution,
rather than gremlin controlling it through the java api's. 

This is a big deal to do. You would be “all in” on TinkerPop and someone in your organization would be the TinkerPop-expert. In essence, you would be maintaining the implementation of a virtual machine. We would be best of friends and that is not a good thing. 

As a side: It would be nice if you were in C (vs. Go) as then your Gremlin VM would be able to be leveraged by others — e.g. ArangoDB, MongoDB, etc. I don’t know of any graph databases that are implemented in Go. ?

Example: Say we need to find all likes and dislikes of a person, dgraph shards the graph by predicates rather then vertices, so likes
and dislikes of a person can be found in parallel. It would be tedious to implement all these vendor strategies in gremlin(option 3 and 4)
(It's a duplication of code) and i am not sure whether all optimisations could be implemented in gremlin.

You can always implement optimizations in Gremlin. Gremlin’s compiler model (TraversalStrategies) is very flexible and you should be able to accomplish as much threading/bulking as you want. 

Good luck,
Marko.





Janardhan Reddy

unread,
Nov 2, 2016, 9:28:25 AM11/2/16
to Gremlin-users
Hello,

Thanks for your inputs. Can you please elaborate why you would choose 3 vs 1 and 4 vs 2.

Hello,


-- 
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/c83e0ad2-0051-4daa-e1f5-e1ebaf151141%40gmail.com.
For more options, visit https://groups.google.com/d/optout.

Stephen Mallette

unread,
Nov 2, 2016, 10:58:09 AM11/2/16
to Gremlin-users
>  It would be nice if you were in C (vs. Go) as then your Gremlin VM would be able to be leveraged by others — e.g. ArangoDB, MongoDB, etc. I don’t know of any graph databases that are implemented in Go. ?

Well, DGraph is written in Go which is presume is why Janardhan is interested in the possibilities. Cayley is the other that i can think of.

Hello,


-- 
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/c83e0ad2-0051-4daa-e1f5-e1ebaf151141%40gmail.com.
For more options, visit https://groups.google.com/d/optout.

-- 
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-users+unsubscribe@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-users+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/91DE913E-450C-4921-B56E-681E7F37A8F0%40gmail.com.

Janardhan Reddy

unread,
Nov 2, 2016, 3:02:47 PM11/2/16
to Gremlin-users
Hi,

I get it  that in general option 3 and 4 are better than option 1 and 2.  
Graphql doesn't provide a way to specify graph queries e.g we can't express the shortest path from one node to another in graphql. With option 1 and 2 we would loose some of the features of gremlin itself like we can't specify some steps like until and simplepath in graphql.

Option 4 would be too much tedious if we want to support full gremlin machine(side effects, match, traversal, etc). 
Option 3 seems to less tedious than option 3 but we need to figure out bulk strategies. - It won't exactly be a network call right ?, but calls between gremlin server and dgraph would be through java-go binding since the language variants and gremlin console pass would pass the whole gremlin bytecode to gremlin server.

I was just wondering since option 3  was implemented by most graph vendors, have they implemented better strategies for bulk operations(non jvm graph vendors). For jvm vendors there won't be any overhead but for non jvm vendors the java - non-jvm language linkage would be slower and there is a need for bulk operations.
Hello,

To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-user...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-user...@googlegroups.com.

Janardhan Reddy

unread,
Nov 2, 2016, 3:12:33 PM11/2/16
to Gremlin-users
 I get it  that in general option 3 and 4 are better than option 1 and 2.  
Graphql doesn't provide a way to specify graph queries e.g we can't express the shortest path from one node to another in graphql. With option 1 and 2 we would loose some of the features of gremlin itself like we can't specify some steps like until and simplepath in graphql.

Option 4 would be too much tedious if we want to support full gremlin machine(side effects, match, traversal, etc). 
Option 3 seems to less tedious than option 3 but we need to figure out bulk strategies.

I was just wondering since option 3  was implemented by most graph vendors, have they implemented better strategies for bulk operations.

What do you suggest between option 3 and 4.Even with option 4 we would need similar optimisations right ?

Marko Rodriguez

unread,
Nov 2, 2016, 3:41:01 PM11/2/16
to gremli...@googlegroups.com
Hi,

I was just wondering since option 3  was implemented by most graph vendors, have they implemented better strategies for bulk operations.

Apache TinkerPop’s strategies are all about optimizing to what is possible with TinkerPop API. 

Vendors will implement ProviderOptimizationStrategies. These will take advantage of unique features of the vendor’s graph system. A classic examples is index lookups. g.V().has(‘name’,’marko’) can be a single index call as opposed to iterating out all vertices and filtering one by one.

Sqlg turns a series of steps into a single operation.
Titan does this as well.
TinkerGraph does this for global index lookups.
Neo4jGraph does this for global index lookups.
SparkGraphComputer does this for non-traversing reducing traversals — e.g. g.V().count(), g.V().groupCount().by(label), etc.
…
in short, every vendor will ultimately implement custom strategies.

Once you have the TinkerPop structure API implemented, everything works (full test suite passes, you are happy). From there, you will nit pick and say stuff like: “huh, if I see a out().has(“name”,”marko”), I can do that as a single call cause I will already know the properties on the adjacent vertex.” There you will have a strategy that folds such patterns into a custom step that does as you need. Run the test suite again. Everything passes. Sweet. So slowly, you can just add more and more tricks to make Gremlin’s integration with your system all the more optimized.

See:

HTH,
Marko.


Janardhan Reddy

unread,
Nov 3, 2016, 1:37:07 AM11/3/16
to Gremlin-users
Hi,

Is it possible to give some part of traversal control to vendor as in say SimplePath(), since dgraph stores the data shared by edge label, the simple path can be retrieved by DGraph with a single network call.(The whole data might lie on a single machine).

Marko Rodriguez

unread,
Nov 3, 2016, 9:12:38 AM11/3/16
to gremli...@googlegroups.com
Hello,

Is it possible to give some part of traversal control to vendor as in say SimplePath(), since dgraph stores the data shared by edge label, the simple path can be retrieved by DGraph with a single network call.(The whole data might lie on a single machine).

Yes! That is the point. Implement the TinkerPop structure API first and Gremlin will just work. Stage 1 complete. At this point, Apache TinkerPop will have full control of the execution. However, when you want to optimize, you write a strategy. This will delegate certain aspects of the traversal’s execution over to the vendor. For things the vendor can do better than Apache TinkerPop, it does. For things that the vendor can’t do better, Apache TinkerPop handles it.

I don’t know how many strategies Sqlg has, but I believe DSEGraph has 3 — one for global index use (g.V().has(’name’,’marko’)), one for vertex-centric index use (outE().has(‘time’,gt(2001))), and one for batch “get” of vertex properties.

HTH,
Marko.


Janardhan Reddy

unread,
Nov 6, 2016, 11:33:34 PM11/6/16
to Gremlin-users
Hello Marko,

We want to jumpstart to make Dgraph tinkerpop enabled. I was wondering if you are available for a chat with Dgraph team to clarify our doubts and to finalise a way to integrate tinkerpop in Dgraph.

Thanks

Marko Rodriguez

unread,
Nov 7, 2016, 11:29:34 AM11/7/16
to gremli...@googlegroups.com
Hello,

I ping’d you in the other email you sent me.

However, now that I think about it — perhaps, we open up a Google Hangout for anyone interested in hearing the discussion. That is, we discuss how to get Dgraph set up with TinkerPop and anyone can listen in… ?

Marko.

pieter-gmail

unread,
Nov 7, 2016, 12:12:13 PM11/7/16
to gremli...@googlegroups.com
Hi,

I for one am interested in this discussion.

Thanks
Pieter


On 07/11/2016 18:29, Marko Rodriguez wrote:
> Hello,
>
> I ping’d you in the other email you sent me.
>
> However, now that I think about it — perhaps, we open up a Google
> Hangout for anyone interested in hearing the discussion. That is, we
> discuss how to get Dgraph set up with TinkerPop and anyone can listen
> in… ?
>
> Marko.
>
> http://markorodriguez.com
>
>
>
>> On Nov 6, 2016, at 9:33 PM, Janardhan Reddy
>> <annapareddyj...@gmail.com
>>> <http://tinkerpop.apache.org/docs/current/reference/#traversalstrategy>
>>> http://tinkerpop.apache.org/providers.html
>>> <http://tinkerpop.apache.org/providers.html>
>>> http://tinkerpop.apache.org/gremlin.html
>>>> <http://markorodriguez.com/>
>>>>> <http://www.datastax.com/dev/blog/the-benefits-of-the-gremlin-graph-traversal-machine>
>>>>> https://markorodriguez.com/2011/08/03/on-the-nature-of-pipes/
>>>>> <https://markorodriguez.com/2011/08/03/on-the-nature-of-pipes/> (very
>>>>> old, but nice explanatory graphics)
>>>>> https://markorodriguez.com/2011/04/19/local-and-distributed-traversal-engines/
>>>>> <https://markorodriguez.com/2011/04/19/local-and-distributed-traversal-engines/>
>>>>> https://markorodriguez.com/2011/02/08/property-graph-algorithms/
>>>>> <https://markorodriguez.com/2011/02/08/property-graph-algorithms/> (again,
>>>>> very old, but concepts)
>>>>> If you are interested in such a path, I
>>>>> would be willing to write a
>>>>> Gremlin(mini-mini) in Python (lets say)
>>>>> and write a blog post about it so people
>>>>> can learn how to get started.
>>>>>
>>>>> Finally, if you do go down this route, I
>>>>> strongly advise to use the same
>>>>> terminology and structure as Apache
>>>>> Tinkerpop’s Gremlin machine. With a direct
>>>>> mapping to Go, discussions, advances, etc.
>>>>> will be so much easier…
>>>>>
>>>>> HTH,
>>>>> Marko.
>>>>>
>>>>> http://markorodriguez.com
>>>>> <http://markorodriguez.com/>
>>>>>
>>>>>
>>>>>
>>>>>> On Nov 1, 2016, at 12:21 PM, pieter-gmail
>>>>>> <pieter...@gmail.com <http://gmail.com/>>
>>>>>>> <http://markorodriguez.com/>
>>>>>>>> <http://googlegroups.com/>.
>>>>>>>> To view this discussion on
>>>>>>>> the web
>>>>>>>> visit https://groups.google.com/d/msgid/gremlin-users/a81ee7da-f719-473a-b85c-4e495ad7a5c4%40googlegroups.com
>>>>>>>> <https://groups.google.com/d/msgid/gremlin-users/a81ee7da-f719-473a-b85c-4e495ad7a5c4%40googlegroups.com>
>>>>>>>> <https://groups.google.com/d/msgid/gremlin-users/a81ee7da-f719-473a-b85c-4e495ad7a5c4%40googlegroups.com?utm_medium=email&utm_source=footer
>>>>>>>> <https://groups.google.com/d/msgid/gremlin-users/a81ee7da-f719-473a-b85c-4e495ad7a5c4%40googlegroups.com?utm_medium=email&utm_source=footer>>.
>>>>>>>> For more options,
>>>>>>>> visit https://groups.google.com/d/optout
>>>>>>>> <https://groups.google.com/d/optout>
>>>>>>>> <https://groups.google.com/d/optout
>>>>>>>> <https://groups.google.com/d/optout>>.
>>>>>>>
>>>>>>> --
>>>>>>> You received this message because you
>>>>>>> are subscribed to the Google
>>>>>>> Groups "Gremlin-users" group.
>>>>>>> To unsubscribe from this group and stop
>>>>>>> receiving emails from it, send
>>>>>>> an email
>>>>>>> to gremlin-user...@googlegroups.com
>>>>>>> <http://googlegroups.com/>
>>>>>>> <mailto:gremlin-user...@googlegroups.com>.
>>>>>>> To view this discussion on the web visit
>>>>>>> https://groups.google.com/d/msgid/gremlin-users/3c2fcba2-04b5-4e71-b9bd-778f0f8841c5%40googlegroups.com
>>>>>>> <https://groups.google.com/d/msgid/gremlin-users/3c2fcba2-04b5-4e71-b9bd-778f0f8841c5%40googlegroups.com>
>>>>>>> <https://groups.google.com/d/msgid/gremlin-users/3c2fcba2-04b5-4e71-b9bd-778f0f8841c5%40googlegroups.com?utm_medium=email&utm_source=footer
>>>>>>> <https://groups.google.com/d/msgid/gremlin-users/3c2fcba2-04b5-4e71-b9bd-778f0f8841c5%40googlegroups.com?utm_medium=email&utm_source=footer>>.
>>>>>>> For more options,
>>>>>>> visit https://groups.google.com/d/optout
>>>>>>> <https://groups.google.com/d/optout>.
>>>>>>
>>>>>> --
>>>>>> You received this message because you are
>>>>>> subscribed to the Google Groups
>>>>>> "Gremlin-users" group.
>>>>>> To unsubscribe from this group and stop
>>>>>> receiving emails from it, send an email
>>>>>> to gremlin-user...@googlegroups.com
>>>>>> <http://googlegroups.com/>.
>>>>>> To view this discussion on the web
>>>>>> visit https://groups.google.com/d/msgid/gremlin-users/c83e0ad2-0051-4daa-e1f5-e1ebaf151141%40gmail.com
>>>>>> <https://groups.google.com/d/msgid/gremlin-users/c83e0ad2-0051-4daa-e1f5-e1ebaf151141%40gmail.com>.
>>>>>> For more options,
>>>>>> visit https://groups.google.com/d/optout
>>>>>> <https://groups.google.com/d/optout>.
>>>>>
>>>>>
>>>>> --
>>>>> You received this message because you are
>>>>> subscribed to the Google Groups
>>>>> "Gremlin-users" group.
>>>>> To unsubscribe from this group and stop
>>>>> receiving emails from it, send an email
>>>>> to gremlin-user...@googlegroups.com.
>>>>> To view this discussion on the web
>>>>> visit https://groups.google.com/d/msgid/gremlin-users/6e39b325-aebc-447e-8bc3-793de0aaa481%40googlegroups.com
>>>>> <https://groups.google.com/d/msgid/gremlin-users/6e39b325-aebc-447e-8bc3-793de0aaa481%40googlegroups.com?utm_medium=email&utm_source=footer>.
>>>>> For more options,
>>>>> visit https://groups.google.com/d/optout
>>>>> <https://groups.google.com/d/optout>.
>>>>
>>>>
>>>> --
>>>> You received this message because you are
>>>> subscribed to the Google Groups "Gremlin-users"
>>>> group.
>>>> To unsubscribe from this group and stop
>>>> receiving emails from it, send an email
>>>> to gremlin-user...@googlegroups.com.
>>>> To view this discussion on the web
>>>> visit https://groups.google.com/d/msgid/gremlin-users/91DE913E-450C-4921-B56E-681E7F37A8F0%40gmail.com
>>>> <https://groups.google.com/d/msgid/gremlin-users/91DE913E-450C-4921-B56E-681E7F37A8F0%40gmail.com?utm_medium=email&utm_source=footer>.
>>>>
>>>>
>>>> For more options,
>>>> visit https://groups.google.com/d/optout
>>>> <https://groups.google.com/d/optout>.
>>>>
>>>>
>>>>
>>>> --
>>>> You received this message because you are subscribed to the
>>>> Google Groups "Gremlin-users" group.
>>>> To unsubscribe from this group and stop receiving emails
>>>> from it, send an email to gremlin-user...@googlegroups.com
>>>> <http://googlegroups.com/>.
>>>> To view this discussion on the web
>>>> visit https://groups.google.com/d/msgid/gremlin-users/4625e2fb-bf25-4174-8a8b-b44200a2385a%40googlegroups.com
>>>> <https://groups.google.com/d/msgid/gremlin-users/4625e2fb-bf25-4174-8a8b-b44200a2385a%40googlegroups.com?utm_medium=email&utm_source=footer>.
>>>> For more options, visit https://groups.google.com/d/optout
>>>> <https://groups.google.com/d/optout>.
>>>
>>>
>>> --
>>> You received this message because you are subscribed to the
>>> Google Groups "Gremlin-users" group.
>>> To unsubscribe from this group and stop receiving emails from
>>> it, send an email to gremlin-user...@googlegroups.com
>>> <http://googlegroups.com/>.
>>> To view this discussion on the web
>>> visit https://groups.google.com/d/msgid/gremlin-users/fc4e660e-cd7b-4c39-bbb1-07bfb42ce5fb%40googlegroups.com
>>> <https://groups.google.com/d/msgid/gremlin-users/fc4e660e-cd7b-4c39-bbb1-07bfb42ce5fb%40googlegroups.com?utm_medium=email&utm_source=footer>.
>>> For more options, visit https://groups.google.com/d/optout
>>> <https://groups.google.com/d/optout>.
>>
>>
>> --
>> You received this message because you are subscribed to the Google
>> Groups "Gremlin-users" group.
>> To unsubscribe from this group and stop receiving emails from it,
>> send an email to gremlin-user...@googlegroups.com
>> <mailto:gremlin-user...@googlegroups.com>.
>> To view this discussion on the web
>> visit https://groups.google.com/d/msgid/gremlin-users/6c0e3a0e-ddd6-455c-b5ca-de69056395c4%40googlegroups.com
>> <https://groups.google.com/d/msgid/gremlin-users/6c0e3a0e-ddd6-455c-b5ca-de69056395c4%40googlegroups.com?utm_medium=email&utm_source=footer>.
>> For more options, visit https://groups.google.com/d/optout.
>
> --
> You received this message because you are subscribed to the Google
> Groups "Gremlin-users" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to gremlin-user...@googlegroups.com
> <mailto:gremlin-user...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/gremlin-users/9723BC7E-4014-4EBA-83C6-A147E255E8F9%40gmail.com
> <https://groups.google.com/d/msgid/gremlin-users/9723BC7E-4014-4EBA-83C6-A147E255E8F9%40gmail.com?utm_medium=email&utm_source=footer>.

Jason Plurad

unread,
Nov 7, 2016, 12:44:01 PM11/7/16
to gremli...@googlegroups.com
+1 on the Hangout, great suggestion Marko
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/74866a98-b3f7-d064-3900-4a9ccde7c29b%40gmail.com.

Janardhan Reddy

unread,
Nov 7, 2016, 1:26:24 PM11/7/16
to Gremlin-users
Hello,

Thanks for accepting my request. Would you be available between 18:00 to  20:00 (mexico time). (GMT 00:00 - 02:00) say tomorrow..  I couldn't find any other time slot which works since i am based in GMT+5:30 timezone, dgraph team is sydney and you are based in US.

On Monday, 7 November 2016 23:14:01 UTC+5:30, Jason Plurad wrote:
+1 on the Hangout, great suggestion Marko
On Mon, Nov 7, 2016 at 12:12 PM pieter-gmail <pieter...@gmail.com> wrote:
Hi,

I for one am interested in this discussion.

Thanks
Pieter


On 07/11/2016 18:29, Marko Rodriguez wrote:
> Hello,
>
> I ping’d you in the other email you sent me.
>
> However, now that I think about it — perhaps, we open up a Google
> Hangout for anyone interested in hearing the discussion. That is, we
> discuss how to get Dgraph set up with TinkerPop and anyone can listen
> in… ?
>
> Marko.
>
> http://markorodriguez.com
>
>
>
>> On Nov 6, 2016, at 9:33 PM, Janardhan Reddy
>> <annapareddyj...@gmail.com
>>>>>>>                         <mailto:gremlin-users+unsub...@googlegroups.com>.

>> To view this discussion on the web
>> visit https://groups.google.com/d/msgid/gremlin-users/6c0e3a0e-ddd6-455c-b5ca-de69056395c4%40googlegroups.com
>> <https://groups.google.com/d/msgid/gremlin-users/6c0e3a0e-ddd6-455c-b5ca-de69056395c4%40googlegroups.com?utm_medium=email&utm_source=footer>.
>> For more options, visit https://groups.google.com/d/optout.
>
> --
> You received this message because you are subscribed to the Google
> Groups "Gremlin-users" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to gremlin-user...@googlegroups.com

Austin Harris

unread,
Nov 7, 2016, 1:42:25 PM11/7/16
to Gremlin-users
I would listen as well.


On Monday, November 7, 2016 at 9:29:34 AM UTC-7, Marko A. Rodriguez wrote:
Hello,

Stephen Mallette

unread,
Nov 8, 2016, 6:50:35 AM11/8/16
to Gremlin-users
I would have been interested in joining, but can't make that time. Depending on the direction this goes, it will be interesting to see how the test suite will have to evolve to support non-jvm languages. We had some discussion on this with the dev list, but no real action has been taken yet. if Gremlin, whether a GLV or a machine, is to be implemented off the JVM, the test suite needs to be better suited to work in any language (preferably in some native way). I'd suggested the use of Cucumber based tests to help support this need. I'd recently learned that Cypher takes that same approach with their Technology Compliance Kit:




--
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-users+unsubscribe@googlegroups.com.
 To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/7447f81f-6fbd-4aa5-a5b6-93ba19b11458%40googlegroups.com.

Marko Rodriguez

unread,
Nov 8, 2016, 7:00:56 AM11/8/16
to gremli...@googlegroups.com
Hello,

I can do Friday morning of this week.

Marko.

Marko Rodriguez

unread,
Nov 8, 2016, 7:15:21 AM11/8/16
to gremli...@googlegroups.com
I don’t know what Cucumber is but here is how I see it.

If a language can generate Bytecode (which it needs to), then gremlin-tests will be able to test it. However, this means something like this:

Gremlin language variant -> Bytecode -> GremlinServer -> gremlin-test.

It would mean standing up some interface (e.g. GremlinServer) between the Gremlin language variant and the Apache TinkerPop Gremlin traversal machine. Seems heavy weight. ??

Marko.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/CAA-H439fW63cznp3n0xHNM17YE86kZb9N37GOe_Pprgaqa%2BH5w%40mail.gmail.com.

Stephen Mallette

unread,
Nov 8, 2016, 7:42:48 AM11/8/16
to Gremlin-users
Yeah - it could be done that way, but Including Gremlin Server as part of the test suite is what I think we should try to avoid in the long run. Ultimately, it seems that we should have a "Gremlin Specification" that was independent of any particular language or vm. Cucumber would give us a way to do that. Obviously, it's a non-trivial shift in how we do things, but I think that if we'd envisioned GLVs from the start of TinkerPop 3.0 we'd probably have gone with this style of thinking. We unfortunately now have a lot invested in the java test suite and but it's awkward to use off the JVM. In a sense it's even semi-awkward with JVM languages - bytecode reduces that a bit, i suppose, but if we want to see Gremlin popping up in lots of different languages I think that we'll need a more friendly, low impact test suite that can validate the more abstract notion of a "Gremlin Specification" than the specific implementation on the jvm.

--
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-users+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/201A8FB3-0EDC-4273-B877-07AE9C4DAA66%40gmail.com.

pieter-gmail

unread,
Nov 8, 2016, 7:55:40 AM11/8/16
to gremli...@googlegroups.com
A long time ago I wrote a implementation of TinkerPop 2 on lmdb (a
transactional c database) via JNDI.
It was fast but no where near as fast as doing the traversals in C on
the db. The killer being the latency going through the JNDI layer
millions of times.
That made me of the opinion that the query language really ought to be
written close to the db in its technology.

The same holds for Sqlg, a version that is written in postgres PL/pgSql
or via some Postgresql extension written in C will far outperform doing
it via jdbc.

Ideally the db must accept full gremlin strings parse it and return the
result with no latency or memory issues.

So a TinkerPop implementation written Go or C I'd say is first price. A
huge job to be sure but still.

This means TinkerPop will be a java reference implementation rather than
thee implementation with an implementation independent test/semantic
specification.

Cheers
Pieter

On 08/11/2016 14:15, Marko Rodriguez wrote:
> I don’t know what Cucumber is but here is how I see it.
>
> If a language can generate Bytecode (which it needs to), then
> gremlin-tests will be able to test it. However, this means something
> like this:
>
> Gremlin language variant -> Bytecode -> GremlinServer -> gremlin-test.
>
> It would mean standing up some interface (e.g. GremlinServer) between
> the Gremlin language variant and the Apache TinkerPop Gremlin
> traversal machine. Seems heavy weight. ??
>
> Marko.
>
> http://markorodriguez.com
>
>
>
>> On Nov 8, 2016, at 4:50 AM, Stephen Mallette <spmal...@gmail.com
>> <mailto:spmal...@gmail.com>> wrote:
>>
>> I would have been interested in joining, but can't make that time.
>> Depending on the direction this goes, it will be interesting to see
>> how the test suite will have to evolve to support non-jvm languages.
>> We had some discussion on this with the dev list, but no real action
>> has been taken yet. if Gremlin, whether a GLV or a machine, is to be
>> implemented off the JVM, the test suite needs to be better suited to
>> work in any language (preferably in some native way). I'd suggested
>> the use of Cucumber based tests to help support this need. I'd
>> recently learned that Cypher takes that same approach with their
>> Technology Compliance Kit:
>>
>> https://github.com/opencypher/openCypher/tree/master/tck
>>
>>
>>
>> On Mon, Nov 7, 2016 at 1:22 PM, Austin Harris
>> <austin....@gmail.com <mailto:austin....@gmail.com>> wrote:
>>
>> I would listen as well.
>>
>> On Monday, November 7, 2016 at 9:29:34 AM UTC-7, Marko A.
>> Rodriguez wrote:
>>
>> Hello,
>>
>> I ping’d you in the other email you sent me.
>>
>> However, now that I think about it — perhaps, we open up a
>> Google Hangout for anyone interested in hearing the
>> discussion. That is, we discuss how to get Dgraph set up with
>> TinkerPop and anyone can listen in… ?
>>
>> Marko.
>>
>> http://markorodriguez.com <http://markorodriguez.com/>
>>
>>
>>
>>
>>
>> --
>> You received this message because you are subscribed to the
>> Google Groups "Gremlin-users" group.
>> To unsubscribe from this group and stop receiving emails from it,
>> send an email to gremlin-user...@googlegroups.com
>> <mailto:gremlin-user...@googlegroups.com>.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/gremlin-users/7447f81f-6fbd-4aa5-a5b6-93ba19b11458%40googlegroups.com
>> <https://groups.google.com/d/msgid/gremlin-users/7447f81f-6fbd-4aa5-a5b6-93ba19b11458%40googlegroups.com?utm_medium=email&utm_source=footer>.
>>
>>
>> For more options, visit https://groups.google.com/d/optout
>> <https://groups.google.com/d/optout>.
>>
>>
>>
>> --
>> You received this message because you are subscribed to the Google
>> Groups "Gremlin-users" group.
>> To unsubscribe from this group and stop receiving emails from it,
>> send an email to gremlin-user...@googlegroups.com
>> <mailto:gremlin-user...@googlegroups.com>.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/gremlin-users/CAA-H439fW63cznp3n0xHNM17YE86kZb9N37GOe_Pprgaqa%2BH5w%40mail.gmail.com
>> <https://groups.google.com/d/msgid/gremlin-users/CAA-H439fW63cznp3n0xHNM17YE86kZb9N37GOe_Pprgaqa%2BH5w%40mail.gmail.com?utm_medium=email&utm_source=footer>.
>> For more options, visit https://groups.google.com/d/optout.
>
> --
> You received this message because you are subscribed to the Google
> Groups "Gremlin-users" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to gremlin-user...@googlegroups.com
> <mailto:gremlin-user...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/gremlin-users/201A8FB3-0EDC-4273-B877-07AE9C4DAA66%40gmail.com
> <https://groups.google.com/d/msgid/gremlin-users/201A8FB3-0EDC-4273-B877-07AE9C4DAA66%40gmail.com?utm_medium=email&utm_source=footer>.

Marko Rodriguez

unread,
Nov 8, 2016, 8:28:02 AM11/8/16
to gremli...@googlegroups.com
Hi,

So we have two discussions here:

1. Gremlin language test suite.
2. Gremlin machine test suite.

I think Stephen is talking more about the Gremlin language test suite and Pieter, the Gremlin machine test suite.

I agree that we need to make it clear that Apache TinkerPop’s Gremlin-Java (language/machine) are reference implementations. We have made that clear for the language, but not so much for the machine.

Again, I don’t know what Cucumber is all about (I Google’d and didn’t grok the examples), but I know that we can definitely make the ProcessStandard/ComputerSuite language agnostic (that is, it doesn’t have to go through TinkerPop’s Gremlin-Java machine implementation much like it currently doesn’t have to go through the Gremlin-Java language implementation). If we can abstract away the machine, then the test suite interface is basically:

Bytecode -> Test[GremlinTraversalMachine]

As an important side. We can get rid of the gremlin-groovy-tests GroovyStandard/ComputerSuite now that we have Translator which translates Bytecode into a language (identical to reverse engineering bytecode to source code). This is how we are able to test Gremlin-Python without having having to explicitly write out all the tests in Python. We should gut the Groovy language tests and use GroovyTranslator. From there, all we would need to do is turn the Gremlin-Java tests (ProcessStandard/ComputerSuite) into ?Cucumber? tests. This then begs the question — is a validation of the language a validation of the machine? Meaning, if the test suite takes both Bytecode (input) + machine (processor), then if it passes, can we say the machine implementation is semantically correct?

Marko.

To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/44482209-dfc9-8405-0988-0795d9d1dcb0%40gmail.com.

Janardhan Reddy

unread,
Nov 8, 2016, 8:36:46 AM11/8/16
to Gremlin-users
Hi Marko,

Since Dgraph's team is based in sydney, the ideal time to meet would be somewhere between 15:00 to 18:00(CST time). It would be midnight for me though considering i am located in GMT+5:30 time zone, but it can't be avoided. Are you available between 15:00 to 18:00 (CST time) rather than morning.

   To view this discussion on the web visit
   https://groups.google.com/d/msgid/gremlin-users/7447f81f-6fbd-4aa5-a5b6-93ba19b11458%40googlegroups.com
   <https://groups.google.com/d/msgid/gremlin-users/7447f81f-6fbd-4aa5-a5b6-93ba19b11458%40googlegroups.com?utm_medium=email&utm_source=footer>.


   For more options, visit https://groups.google.com/d/optout
   <https://groups.google.com/d/optout>.



-- 
You received this message because you are subscribed to the Google
Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to gremlin-user...@googlegroups.com

-- 
You received this message because you are subscribed to the Google
Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send
an email to gremlin-user...@googlegroups.com

Marko Rodriguez

unread,
Nov 8, 2016, 9:17:15 AM11/8/16
to gremli...@googlegroups.com
Hi,

That would be 2pm for me (I’m MST — +1 California). That late into Friday would not be fun. How about we do Thursday at 15:00 CST? 

Next: 3 hours seems intense. Lets shoot for an hour. If it bleeds over, great or if we need to do another call, we can do a Part 2 the following week.

Does that work for everyone?

Marko.


Janardhan Reddy

unread,
Nov 8, 2016, 10:02:51 AM11/8/16
to Gremlin-users
I didn't mean 3 hours. I was just suggesting that we could have a call anytime in between those 3 hours :)

Thursday 15:00 CST sounds good to me.

Janardhan Reddy

unread,
Nov 8, 2016, 10:10:37 AM11/8/16
to Gremlin-users
Here's the hangout link(We can join here whenever we are meeting)

https://hangouts.google.com/hangouts/_/dgraph.io/annapareddyjana?authuser=0

Stephen Mallette

unread,
Nov 8, 2016, 6:15:23 PM11/8/16
to Gremlin-users
I think I'm talking about both discussions - language and machine test suites. I've thought less about what the machine test suite would look like, but it seems related in my mind, no?


   To view this discussion on the web visit
   https://groups.google.com/d/msgid/gremlin-users/7447f81f-6fbd-4aa5-a5b6-93ba19b11458%40googlegroups.com
   <https://groups.google.com/d/msgid/gremlin-users/7447f81f-6fbd-4aa5-a5b6-93ba19b11458%40googlegroups.com?utm_medium=email&utm_source=footer>.


   For more options, visit https://groups.google.com/d/optout
   <https://groups.google.com/d/optout>.



-- 
You received this message because you are subscribed to the Google
Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it,

-- 
You received this message because you are subscribed to the Google
Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send

-- 
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-users+unsubscribe@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-users+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/C1ADD237-C1D2-47CE-A52D-A27F5B0B17BD%40gmail.com.

Janardhan Reddy

unread,
Nov 10, 2016, 1:51:47 AM11/10/16
to Gremlin-users
Hello,

Just a reminder that we are having a hangout at Thursday 15:00 CST (Today). Anyone interested can join.

https://hangouts.google.com/hangouts/_/dgraph.io/annapareddyjana?authuser=0

   send an email to gremlin-user...@googlegroups.com

   <mailto:gremlin-users+unsub...@googlegroups.com>.
   To view this discussion on the web visit
   https://groups.google.com/d/msgid/gremlin-users/7447f81f-6fbd-4aa5-a5b6-93ba19b11458%40googlegroups.com
   <https://groups.google.com/d/msgid/gremlin-users/7447f81f-6fbd-4aa5-a5b6-93ba19b11458%40googlegroups.com?utm_medium=email&utm_source=footer>.


   For more options, visit https://groups.google.com/d/optout
   <https://groups.google.com/d/optout>.



-- 
You received this message because you are subscribed to the Google
Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it,

-- 
You received this message because you are subscribed to the Google
Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send

-- 
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-user...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-user...@googlegroups.com.

Jason Plurad

unread,
Nov 15, 2016, 11:13:47 AM11/15/16
to Gremlin-users
I wasn't able to attend. Would Janardhan or anybody else like to share a summary post of this meeting? Thanks.

-- Jason


   To view this discussion on the web visit
   https://groups.google.com/d/msgid/gremlin-users/7447f81f-6fbd-4aa5-a5b6-93ba19b11458%40googlegroups.com
   <https://groups.google.com/d/msgid/gremlin-users/7447f81f-6fbd-4aa5-a5b6-93ba19b11458%40googlegroups.com?utm_medium=email&utm_source=footer>.


   For more options, visit https://groups.google.com/d/optout
   <https://groups.google.com/d/optout>.



-- 
You received this message because you are subscribed to the Google
Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it,

-- 
You received this message because you are subscribed to the Google
Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send
an email to gremlin-user...@googlegroups.com

-- 
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/44482209-dfc9-8405-0988-0795d9d1dcb0%40gmail.com.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
 To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/C1ADD237-C1D2-47CE-A52D-A27F5B0B17BD%40gmail.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Gremlin-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gremlin-users+unsubscribe@googlegroups.com.
 To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/d97ffea7-eec7-4115-bb3f-a35bb4d46c1d%40googlegroups.com.

Stephen Mallette

unread,
Nov 16, 2016, 4:06:34 PM11/16/16
to Gremlin-users
Jason, I don't think you missed much that you didn't know already as you already know how TinkerPop works and how it is structured. Most of the discussion sorta rolled around that and how to do so to get some quick wins in terms of "hey - it's working!" Obviously strategy development was discussed as that is a critical piece to making a provider implementation show off it's capabilities. No big new outcomes like a "Go Gremlin VM" in the current works or anything like that.

On Tue, Nov 15, 2016 at 11:13 AM, Jason Plurad <plu...@gmail.com> wrote:
I wasn't able to attend. Would Janardhan or anybody else like to share a summary post of this meeting? Thanks.

-- Jason
To view this discussion on the web visit https://groups.google.com/d/msgid/gremlin-users/CAOk5v9Muf2a4_bfXvyQ%2BrPSBnOavL1TE9APet_zsOhVr95sF1Q%40mail.gmail.com.
Reply all
Reply to author
Forward
0 new messages