Your drivers are quite messy,any plan to clean them up?

[AlexLuya]

Honestly,at first glance,nobody will believe that a db provider will do that like you guys:totally messy. 1,What drivers are supported officially? 2,What does the driver type mean?

[rotatingJazz]

Too much free time and decided to troll around? That 's what IRC is for, not github issues.

[frankmayer]

@AlexLuya If you'd like to discuss something, which is not directly an issue, then please do so politely at the forums =>(https://groups.google.com/forum/#!forum/arangodb).

If you really wanted to point out an issue, you should state exactly what you think the problem is. Trolling around does not make a valid Github issue.

As for your points:

Honestly,at first glance,nobody will believe that a db provider will do that like you guys:totally messy.

Short version: Honestly... your sentence really does make no sense: totally messy....

Long version: You've got to get your sentences right. What will nobody believe, that a db provider will do, like the guys did? What exactly is totally messy?

1,What drivers are supported officially?

Some drivers and extensions to drivers have been developed by the ArangoDB team members. They are supported either by the members or other users like myself (I am not affiliated with the company behind ArangoDB). I'll leave the "officially supported" part to the team itself to answer.

2,What does the driver type mean? Hmm, some driver types I got off the drivers page: http://www.arangodb.org/api

• Type: Blueprint driver
• Type: ArangoDB Driver for D
• Type: object-document mapper
• Type: Client for Node.js
• Type: PHP ODM for arangoDB
• Type: ArangoDB Adapter for JugglingDB ORM

Driver type just states, as the word says, the type of the driver, like ODM or Blueprint, etc...

There is a 'tbc' type which to be honest, I wouldn't know what that is at the moment, but I must say....:

Really??? That is your problem with the Database Server project?

If you have some real issue to report (and ultimately also propose some solution), then please do so in a polite manner.

[AlexLuya]

Sorry,intended to say it like this to alert you guys to take care of it.Honestly,if needing a document db,user will try mongodb or rethinkdb,if key-store db needed,user will consider redis or memched.Why will I touch arangodb?Because,I need a graph db,and compare to neo4j, orientdb,arangodb is c++ based.And I watched arangodb longer than one year,and even you drivers are not "messy",they are not so good,maybe you guys have not enough resource on it,but from strategy point of view,it worth taking more efforts on it and let users to help you find bugs.

[frankmayer]

@AlexLuya Hmmm, I am not sure I understand what points you are trying to make.

This is an open source project and except for 2 or 3 drivers where team members are involved, the drivers are being written and maintained by the community.

If you believe that some driver is not so good, why don't you contribute to make it better?

(Honestly, I'd like to know which drivers you are referring to, and to what you are comparing them to...)

...and let users to help you find bugs.

Again, it's an open source project and everyone is welcome to help in any way... (providing bug reports, drivers, docs, corrections, etc...). The team is not discouraging anyone to help find bugs...

[AlexLuya]

Ok,make it clear,to make user feels a product is promising: 1,Must have some drivers got supported officially,and must be updated with product; 2,As you guys did currently,too many drivers version coexisted and listed in same page,and most of them is not updated.At first glance,user will think: why?Has most of driver maintainers just gone away?Even,this isn't truth.

[weinberger]

I’d like to make to remarks from a team’s point of view.

Generally, all comments are welcome but the more precise the better. This allows us to dig deeper into them. If I’ve understood you correctly, you are saying that the graph features aren’t supported properly. If that’s correct, please drop us a line describing the issue more precisely.

Interesting to me is your conclusion that Arango is only useful for graphs because of existing solutions in other areas. I’d like to discuss that matter further with you. Is that really true? Mongo, ok, but what has Rethink to offer beyond us? Ok, key-value is supported good enough by Redis, but wouldn’t it be interesting to dodge establishing an extra technology when you already use a document store?

Am 28.10.2013 um 08:22 schrieb Alex Luya notifications@github.com:

Sorry,intended to say it like this to alert you guys to take care of it.Honestly,if needing a document db,user will try mongodb or rethinkdb,if key-store db needed,user will consider redis or memched.Why will I touch arangodb?Because,I need a graph db,and compare to neo4j, orientdb,arangodb is c++ based.And I watched arangodb longer than one year,and even you drivers are not "messy",they are not so good,maybe you guys have not enough resource on it,but from strategy point of view,it worth taking more efforts on it and let users to help you find bugs.

— Reply to this email directly or view it on GitHub.

[AlexLuya]

"Interesting to me is your conclusion that Arango is only useful for graphs",no,that is not what I mean,I just expressed the first feel in some users mind:c++ based graph data is a attractive point.And your argument,maybe, fell into technical circle.I think,as most the important component in a system,user will very,very conservative to choose a db,they or myself will choose one that is being used by most of people,if not such a choose existed and a risk must be token,I will try some candidates that look promising:many people is trying it,bugs got fixed quickly,and api is maintained well,make it easy to try.So,it is not a pure technical problem,but combined some psychologic factors.

[frankmayer]

@AlexLuya I think I understand what you mean.

Let me give you some points for why ArangoDB won my trust.

• When I was evaluating my options needing a document store, my first candidate was MongoDB. Having done my evaluations on MongoDB and I saw some of its drawbacks. While evaluating further, I stumbled upon ArangoDB. It was just at version 1.0 back then, but it looked very promising. What finally won me over, was the character of the team. I Had a few questions on the ArangoDB Google Group, which they answered quickly, thoroughly and it immediately felt "good" (To come back to your last point of "...combined some psychologic factors"). It also gave me ideas to solve some design issues, which I could not have solved with MongoDB alone. To finish up this paragraph... Unfortunately the project that I was evaluating ArangoDB for, was put on hold, but ArangoDB has since added a lot of very important features at a good pace, and I'll definitely be using it for any upcoming project, to involve document store and/or graphs.
• Your point, of a user choosing the most used db is partially true. Yes, most look at how many questions and answers there are on StackOverflow for example, and derive a user base from that, or read some Reports on the most used DB's etc... , but nothing beats a good evaluation of a system. And a lot of Database Servers claim to have some very "wanted" features, but in reality they might not live up to the full user expectations. The thing with ArangoDB, is that the team doesn't claim anything that isn't true. They tell you what exactly ArangoDB can do, what it can't and what it probably will be doing in future. And to me as a developer, that's assuring.
• And to be fair, ArangoDB at least in my eyes, is still a younger project, which with the upcoming version 1.4 will again do a major release, with lots of new and very important features. If MongoDB would not be there (being one of the first very "known" document stores), ArangoDB would have a much wider user base, and a lot more people contributing to it. That being said, ArangoDB 's user base is growing day by day which I find very pleasing :+1:
• Finally on the quality points you make.. the team is very devoted to the project. Medium- and high-priority bugs are most of the time fixed the same day. That's not something you see with most projects out there, unless they are backed by really big companies... and even then, bugs and feature requests get pushed back... The API is growing and is maintained well. Since it's all HTTP, it's easy to try even without having a driver. Use a REST client and start testing. There are a lot of generic REST clients as browser plugins or stand-alone ones, out there. If there is no driver for your language, just get in contact with the maintainer, contribute or write up your own, giving back to the community. The protocol is clearly defined, not a black box.

All in all, the psychological factors, you are referring to, are being met. You just need to look closer :smile:

[AlexLuya]

"You just need to look closer" This is the point,to attract more users to look closer,I think two points blow are very important to this project:

1,Updated,official supported drivers; 2,C++ based graph data is an attractive point,for me or hope more people,and if possible,make it shining.

Hope these suggestions helps.