Follow Slashdot blog updates by subscribing to our blog RSS feed


Forgot your password?

Comment Re:to paraphrase Alice in Dilbert (Score 1) 271

how do you keep an armed private security force loyal?

By carefully selecting them beforehand, and by offering them and their families a place in the hideout.

Give them homes in the shelter and they take over whenever they feel like it

That's the careful selection part. Loyal people still exist in this world, you know?

Comment Re: Clarifying what you really want: (Score 1) 132

Firstly, thanks a lot for your time. This is exactly the kind of exchange I was hoping for (well, that and a small hope someone would post "sure, look here at http://semantic-documentation-...").

Is it absolutely required that the data be embedded in, and tagged within, your documents which are stored in the wiki?

Not absolutely, but I am afraid that if prose and metadata are seperate, they will go out of sync. One thing I like about SMW, for example, is how easy it is to update both at the same time, it comes so naturally that you don't even realise it.

RDFa is RDF in attributes.

Yes, if there were an editor (you already point out there isn't, damn) this would be really cool.

Earlier, you said you figured other people would have wanted to do this before. Of course they have. It's called "The Semantic Web,"

Yes and no. I want something very similar, except with a more narrow focus and thus easier to make possible. I'm currently looking at ArchiMate following someone elses pointers and the advantage of modelling business processes in it as well. Someone in the Netherlands has studied the exact problem I'm trying to solve and published a small paper.

You would not believe the insanely complex data structures necessary to build real, usable databases with only triples.

I believe. The database design for my online game (see footer) is anything but trivial and it has many places where I've encountered exactly this problem of complex relations between entities.

So, I think you would be far better off with a separate database to contain your information. The records in that database could include citations pointing to your documents, in case someone wanted to drill down to the true source of the info.

Yes, I've been thinking along similar lines, except that I would like to embed or something, at least one step better than linking, but that's a UI issue.

Yes, in most documentation you have some prose and then tables or diagrams containing hard data. It would be possible to do it this way, if the systems are well integrated.

You can still incorporate your documentation library, but from the other way around. Instead of storing your information IN your documents and mining it out, you store your information in a purpose-built SYSTEM and then refer to your documents as supporting evidence.

That is another option. What I have learnt from all the answers here put together is that no such solution exists (what a shame. Anyone got VC and looking for a Startup concept?) and that it will either have to be built or put together from existing components. That turns it into an integration issue.

Very interesting answers, thanks. I will follow up on them and talk to others in the company to find the solution that best fits.

Comment Re:Scammers (Score 1) 271

Very smart insight. You should also mention that organized crime has the organisational advantage in not only having a clear command structure, but also one that can operate independently of the public communication networks. Heck, in some countries they have their own mobile networks.

So, it honestly depends on the warehouse owner/manager and the individual connections that the person has.

Depends on the scenario. It makes for a better movie when the breakdown of civilization happens in hours, but from history I would guess that it will take days or weeks, more realistically. That is more than enough to take a few trucks, drive over to the nearest warehouse and simply take everything you want, at gunpoint. Time it right during the breakdown and no police will show up to stop you. You just need to be slightly faster and more organised than other looters.

Comment cheap (Score 2) 271

If you're a billionaire, then 35k is not even worth discussing, even the $3 mio. for an actual space is a minor expense. Easily comparable to what we normal people pay for, say, car insurance. So it's really just that: An "just in case" expense.

Comment Re:Clarifying what you really want: (Score 1) 132

There is existing documentation emerging from the project phase that is largely written in Word and is planned to be added to a wiki or something similar, where it can be better maintained and updated. So every piece of documentation will be touched anyway.

The metadata I am looking for is not about whole documents. I don't give a rats ass about that, if I am looking for the document about encryption protocols, I already know where to find it. I want factlets. From a security perspective, when a network is compromised I want to query the configuration database to find all systems on that subnet, but also the documentation to figure out the destinations of all data flows out of that subnet, for example. When a server fails, I want to understand which other systems that rely on it will likely fail next. When an application is reworked, I want to know which other applications are connected in one way or the other.

All this understanding of the system architecture is usually in the heads of people and hidden somewhere in tomes of documentation and graphs that you need to read and understand. I want that knowledge query-able.

Comment Re:API Dokumentation + Usecases + Glossary ... (Score 1) 132

You are probably right that a knowledge manager would be useful at this stage. And yes, I understand I won't be able to buy a product and press a button and everything is good. But I need to start somewhere, and the right technology choice is usually a good starting point.

Just that I would like to put the metadata into the documentation while it is being written, instead of using some kind of search or analysis in an attempt to extract it later on. It seems stupid to me to not do it right the first time.

Comment Re: Can you stick wiith one solution? (Score 1) 132

As you say this is an IT project It sounds like you need a CMDB that can store relations and metadata. Couple this with confluence for hard documentation and you should be covered.

It is the "couple this" part that brings me here. I want to couple it better than "look here for this and look there for that". If I can integrate these two, I would be covered. Any pointers?

Comment Re: enterprise grade is weasel. (Score 1) 132

That is why I'm looking for a system that understands these problems exist and helps solving them, e.g. by formalizing data input.

That is why I think something like SMW won't cut it, because the system needs structure enforced instead of relying on users always doing the right thing all the time.

It's probably possible to hack something together with SMW and Semantic Forms and some other extensions, I just really wonder if I'm the first person ever who wants structured, query-able documentation?

Comment Re: Atlassian Confluence (Score 1) 132

Why the negativity?

I want to solve a problem, and I have an idea that would improve on the current solution, if I can find something that is not a hack or a DIY. Maybe I will find, maybe I will find out it doesn't exist. I already spent many hours searching and comparing and didn't find, but I already got some good pointers that I can follow for more information. All is good.

Comment Re:RDF/OWL route (Score 1) 132

Yes, the more I read about the subject, the more I tend towards simply using RDF to model the knowledge data, mixed with some kind of wiki or reader to contain the prose parts.

I still think mixing them is more natural to people who write the documentation, and that's what I like SMW for, because you can simply write things like "Server A is on vlan [network:123]" and both human and machine can understand the meaning of that.

But I can't believe a hack to MediaWiki is the best that's available.

1: No code table for op: ++post