The first thing about these articles is to realize that business and government are big proponents. That's why one article about balancing convenience vs. privacy is important. RMS knows this.
That's why a recent NYTimes article about the quants' influence on the financial meltdown quoted Ted Kazcinsky sp? and why an article a few years ago called Why the Future Doesn't Need Us did too.
The second thing to realize is so are consumers of Google and the iPhone.
So all of the kvetching about the use of the term "cloud" really (REALLY) misses the point. Get over it to what it means.
Centralized network terminal computing on central servers is coming and it's going to hit a tipping point that will or may already be affecting your life, depending on your type of business ERP or your own consumer habits.
You might want to look up the short story Manna at MarshallBrain for a dystopian perspective.
Thanks. I saw where someone had made a successful port of KWrite as a demonstration, but have yet to see a schedule or alpha download for Windows testing. et. I'll keep an eye out.
Here are the 5 steps to automation. Figure out what you're doing manually at each step, really figure it out, then automate whatever increment you can. Repeat. Most software / operations today fails at a recognizable stage. Sometimes, you need these steps to occur in parallel, like journaling (gathering) the action someone is doing as they do it, so the journal data can be Categorized, Assigned, and Acted upon by an analyst.
The question to ask always, always, is Why isn't this automated? We're simply in a world moving back to the mainframe. In a mainframe world, redundancy is eliminated.
The obvious solution is to allow access to all information, but set up rules that send alerts on who is searching for the offending words, and disallow anonymous web use, if that even still exists.
This is a very interesting question and I'd mod you up if I could.
We have put in some minimal workflow steps in our installation but now we are looking at (very expensive) commercial tools (tibco, oracle, others) for "BPM".
First, unless you create them, RT doesn't have any structured data points against which you can write workflow rules. Not sure about the other systems mentioned in this thread. So you're limited to queue, Subject line, and body text searches.
That said, you can get part the way down the road using RT: RT already starts by putting tickets into a named queue that can have an email address mapped to the queue or you can simply create a new ticket in a specified queue from within RT. The email interface is quite powerful if you are trying to get processes that have no existing system but email and attachments into some system.
Upon create, you can write rules that assign initial owner based on some data being in the subject. If you're creating tickets manually or from a web form, you can have some structured data in custom RT fields. Then you can write rules based on those data points in PERL.
So, you are kind of at the mercy of how skilled in PERL you are, or in database processing.
It is worth it, though, to write up some sample business processes and see how far you can get with RT, perhaps mated to a wiki, and then relying on persons to follow directions. You can then go after automation and structured data after you really know your work's detailed steps.
hmmm. We've been running it for 2 plus years now, have 100,000 plus tickets, and it's quite fast. We did have to add an index recently when coming back to All Tickets view and many of us have a lot of queues.
I see others have had issues / bad experiences. Our shop has some very experienced Oracle guys and someone who, so far, has been able to make it do everything we've wanted it to using Perl mods. (auto assignment based on subject contents, custom fields, etc.). Maybe other tools are easier when you don't have this kind of support.
We are using it also for project management and in conjunction with Twiki; it's quite effective to create an RT ticket and link it to a wiki page, instead of uploading attachments. This way we end up creating a web FAQ / history and have RT all at once.
We are on a mission at the company where I work to replace all email / attachment based work management with it.
You'd be amazed how far you can push RS using the concepts of owner, status, subject line, journaling, parent child / depends on depended on by tickets, auto-notification, attachments etc. all built in. If you think you need more structured data, you should at least see how far you can get prototyping it first in RT, using its minimal custom fields but also its custom views. Most ERP / CRM don't have the kind of infinite flexibility of workflow you can achieve using the features listed above. They do however have structured data.
First, there's good product design and bad product design, in any field.
Good software comes from good design, based on a combination of study and action, reading up AND doing and evaluating.
For software, it's hard to understand what people really need to do, even when the scope of the software is somewhat limited, like software that supports commercially oriented workflows.
Here are some useful top level terms that can help deal specifically with transaction handling software (think call centers).
Gathering Data (also implies storing data) This means your software getting data from other databases, or emails, or any binary data, or the collection of structured data in forms. Gathering data is a process and history is important so the system knows what effort was involved in the gathering, and to support the idea of rollback to a prior 'state' of data. Systems that keep histories of edits, like Photoshop, have an idea that is rarely applied in transactional systems, at least thay I've seen. Sometimes that is by design, but at other times, it could be permitted in a much better way. Gathering data also means gathering data relationships and lifecyle. How is this related to that? Show me various levels of detail with drill up down and sideways.
Categorizing Data. For transactional systems, the goal is automation. Whenever a person gets involved, it is almost always because gathering or categorizing have failed, or because the next step is outside of the system's awareness--a data gathering issue, like package delivery status. "Business rule" engines and workflow modules live and die by categorization. Q and A systems are all about gathering and categorizing, doing rule outs. There are effective and ineffective methods.
Assigning ownership. Even if the next step in transaction handling is an automated one, System is the owner. Ownership is important for work that involves persons, because commerce needs to understand both quality and cost, which means knowing who did what and how well. Many tranactional systems don't do ownership at all and work gets divided up by querying on other data points like customer, processing status, or product. The form of metadata gathering called Journal or History relies also on owner. Journal and history without owner do not support HR and QA's needs.
Sharing data. This is the whole of all the methods of data presentation, from screens and controls to auto-notifications to any kind of hardware device. Web 2.0 seems to be a lot of hype about Sharing Data and functions (see Acting, below) faster with more persons using a network.
Acting on Data. This is the idea of why the tool you're using was built. Think customer support, purchasing and receiving, transportation, invoicing. The software should ideally work in descending order of desirability like this:
Automated. Relies heavily on categorization and mapping of proper next action
Directed. Think workflow tools, TurboTax interview process, maximo MRO, etc.
Supported by online media with full search, demos, screenshots, tutorials
Supported by non-integrated documentation, wikis, etc.
You learn from your peers and stick notes to your cubicle wall
I think in the article both points of view about why software sucks from the article have some truth. If programmers don't try to do the work the call center person or clerk or whomever is doing from the start and then repeatedly try to do it while it's being built, the wrong thing or the wrong way gets built. Alternately, "Power Users" or "Power Companies" may have too much leverage.
These terms are obviously very broad and have under them all the UI and architectural underpinnings down to how you do multi-select in some field or query on nulls. I wish the author good luck prodding software makers toward better offerings. I suspect like he does, that another article next year won't be out of order.
I must agree with the posters who are saying, Be the Boss. You must do this.
One other thing to do if you can, is have your callers get used to sending you emails and have them do it to an RT server. (http://www.bestpractical.com).
The company where I work is using RT for almost every request / fullfil request scenario that we used to use email for.
It's interestesting to read about the OLPC project and think about what it could to do this kind of cyclical lockin. The OLPC doesn't use intel hardware and it doesn't use Windows.
I think there is some support that the earliest written language had to do with accounting, just so software. Meet the tax man, whether she's queen, the elected officials, whomever. Managing transactions and communications, in a western way, if you choose to view it like that. Reductionism, breaking processes down into their component parts without putting them into context or the idea of their whole particularly well. (Workflow software) Control and bookkeeping.
I think the issue with Google or other search engine is how to do analytics. How do I write a multi-variable where clause? How do I ask a multi-variable question and then hone it or drill into it along one or more parameters, unfolding detail but preserving multiple layers of an outline hierarchy?
So just there is the idea of a different presentation layer, hierarchy and tabular perhaps.
Then, what kind of barriers do I have to getting at the data? Privacy issues? Copyright or patent issues?
If you want to connect two or more points, won't we have to move beyond keyword searches?
Slashdot Mirror
it's relevant...
Why the Future Doesn't Need Us
Generally, we just don't understand all the externalities involved.
Hopefully, they don't lead to catastrophic circumstances.
The first thing about these articles is to realize that business and government are big proponents. That's why one article about balancing convenience vs. privacy is important. RMS knows this.
That's why a recent NYTimes article about the quants' influence on the financial meltdown quoted Ted Kazcinsky sp? and why an article a few years ago called Why the Future Doesn't Need Us did too.
The second thing to realize is so are consumers of Google and the iPhone.
So all of the kvetching about the use of the term "cloud" really (REALLY) misses the point. Get over it to what it means.
Centralized network terminal computing on central servers is coming and it's going to hit a tipping point that will or may already be affecting your life, depending on your type of business ERP or your own consumer habits.
You might want to look up the short story Manna at MarshallBrain for a dystopian perspective.
You may want to think.
If it's not 4F software, it's pricey software.
Four Freedoms. 4F.
If it's not 4F, you lose.
Thanks. I saw where someone had made a successful port of KWrite as a demonstration, but have yet to see a schedule or alpha download for Windows testing. et. I'll keep an eye out.
When will KOffice (supporting ODF) run on Windows?
The question to ask always, always, is Why isn't this automated?
We're simply in a world moving back to the mainframe. In a mainframe world, redundancy is eliminated.
Read Manna.
http://marshallbrain.com/manna1.htm
The obvious solution is to allow access to all information, but set up rules that send alerts on who is searching for the offending words, and disallow anonymous web use, if that even still exists.
Make using the web at all require a login.
This is a very interesting question and I'd mod you up if I could.
We have put in some minimal workflow steps in our installation but now we are looking at (very expensive) commercial tools (tibco, oracle, others) for "BPM".
First, unless you create them, RT doesn't have any structured data points against which you can write workflow rules. Not sure about the other systems mentioned in this thread. So you're limited to queue, Subject line, and body text searches.
That said, you can get part the way down the road using RT:
RT already starts by putting tickets into a named queue that can have an email address mapped to the queue or you can simply create a new ticket in a specified queue from within RT. The email interface is quite powerful if you are trying to get processes that have no existing system but email and attachments into some system.
Upon create, you can write rules that assign initial owner based on some data being in the subject. If you're creating tickets manually or from a web form, you can have some structured data in custom RT fields. Then you can write rules based on those data points in PERL.
So, you are kind of at the mercy of how skilled in PERL you are, or in database processing.
It is worth it, though, to write up some sample business processes and see how far you can get with RT, perhaps mated to a wiki, and then relying on persons to follow directions. You can then go after automation and structured data after you really know your work's detailed steps.
RT can be run in the open web, hosted. BestPractical sells this I think.
Access based solution also != Free.
hmmm.
We've been running it for 2 plus years now, have 100,000 plus tickets, and it's quite fast. We did have to add an index recently when coming back to All Tickets view and many of us have a lot of queues.
I see others have had issues / bad experiences. Our shop has some very experienced Oracle guys and someone who, so far, has been able to make it do everything we've wanted it to using Perl mods. (auto assignment based on subject contents, custom fields, etc.).
Maybe other tools are easier when you don't have this kind of support.
We are using it also for project management and in conjunction with Twiki; it's quite effective to create an RT ticket and link it to a wiki page, instead of uploading attachments. This way we end up creating a web FAQ / history and have RT all at once.
Definitely thumbs up for RT.
We are on a mission at the company where I work to replace all email / attachment based work management with it.
You'd be amazed how far you can push RS using the concepts of owner, status, subject line, journaling, parent child / depends on depended on by tickets, auto-notification, attachments etc. all built in.
If you think you need more structured data, you should at least see how far you can get prototyping it first in RT, using its minimal custom fields but also its custom views.
Most ERP / CRM don't have the kind of infinite flexibility of workflow you can achieve using the features listed above. They do however have structured data.
This is Monsanto...
So will a robot with GPL 3 code be a good idea, so Johnny can make his AIBO do anything he wants?
Has RMS discussed freedom to code and robotics?
Is there any more / less risk with free code than closed code with regard to robotics? Just like any technology in the wrong hands perhaps.
g8orade
Have you looked at Compiere?
RTFM.
This is really poised to be a disrupter.
- Gathering Data (also implies storing data) This means your software getting data from other databases, or emails, or any binary data, or the collection of structured data in forms. Gathering data is a process and history is important so the system knows what effort was involved in the gathering, and to support the idea of rollback to a prior 'state' of data. Systems that keep histories of edits, like Photoshop, have an idea that is rarely applied in transactional systems, at least thay I've seen. Sometimes that is by design, but at other times, it could be permitted in a much better way. Gathering data also means gathering data relationships and lifecyle. How is this related to that? Show me various levels of detail with drill up down and sideways.
- Categorizing Data. For transactional systems, the goal is automation. Whenever a person gets involved, it is almost always because gathering or categorizing have failed, or because the next step is outside of the system's awareness--a data gathering issue, like package delivery status. "Business rule" engines and workflow modules live and die by categorization. Q and A systems are all about gathering and categorizing, doing rule outs. There are effective and ineffective methods.
- Assigning ownership. Even if the next step in transaction handling is an automated one, System is the owner. Ownership is important for work that involves persons, because commerce needs to understand both quality and cost, which means knowing who did what and how well. Many tranactional systems don't do ownership at all and work gets divided up by querying on other data points like customer, processing status, or product. The form of metadata gathering called Journal or History relies also on owner. Journal and history without owner do not support HR and QA's needs.
- Sharing data. This is the whole of all the methods of data presentation, from screens and controls to auto-notifications to any kind of hardware device. Web 2.0 seems to be a lot of hype about Sharing Data and functions (see Acting, below) faster with more persons using a network.
- Acting on Data. This is the idea of why the tool you're using was built. Think customer support, purchasing and receiving, transportation, invoicing. The software should ideally work in descending order of desirability like this:
These terms are obviously very broad and have under them all the UI and architectural underpinnings down to how you do multi-select in some field or query on nulls. I wish the author good luck prodding software makers toward better offerings. I suspect like he does, that another article next year won't be out of order.- Automated. Relies heavily on categorization and mapping of proper next action
- Directed. Think workflow tools, TurboTax interview process, maximo MRO, etc.
- Supported by online media with full search, demos, screenshots, tutorials
- Supported by non-integrated documentation, wikis, etc.
- You learn from your peers and stick notes to your cubicle wall
I think in the article both points of view about why software sucks from the article have some truth. If programmers don't try to do the work the call center person or clerk or whomever is doing from the start and then repeatedly try to do it while it's being built, the wrong thing or the wrong way gets built. Alternately, "Power Users" or "Power Companies" may have too much leverage.I must agree with the posters who are saying, Be the Boss. You must do this.
One other thing to do if you can, is have your callers get used to sending you emails and have them do it to an RT server. (http://www.bestpractical.com).
The company where I work is using RT for almost every request / fullfil request scenario that we used to use email for.
It's interestesting to read about the OLPC project and think about what it could to do this kind of cyclical lockin. The OLPC doesn't use intel hardware and it doesn't use Windows.
Wait and see.
I think there is some support that the earliest written language had to do with accounting, just so software. Meet the tax man, whether she's queen, the elected officials, whomever.
Managing transactions and communications, in a western way, if you choose to view it like that.
Reductionism, breaking processes down into their component parts without putting them into context or the idea of their whole particularly well. (Workflow software)
Control and bookkeeping.
I think the issue with Google or other search engine is how to do analytics.
How do I write a multi-variable where clause?
How do I ask a multi-variable question and then hone it or drill into it along one or more parameters, unfolding detail but preserving multiple layers of an outline hierarchy?
So just there is the idea of a different presentation layer, hierarchy and tabular perhaps.
Then, what kind of barriers do I have to getting at the data? Privacy issues? Copyright or patent issues?
If you want to connect two or more points, won't we have to move beyond keyword searches?
See this link...