I had to recently send a backup of data to a client for them to keep for audit purposes and in case we went bust.
I tarred the data up and encypted it with a random string run through aespipe. I then burnt the data to DVD and posted it to them. Once they confirmed they received the DVD I emailed them the key and told them to keep it safe.
Neither the DVD nor the key are any use on their own. The end user doesn't need any PKI setup at all. There is a readme on the DVD telling how to decrypt it, and they are tasked with keeping the key safe.
Yes they are going after grep. Very much so. They produce a hardware pattern matcher last time I saw (this was back around 1999). This was a chip on a card with a bunch of DIMMs on it. The theory was several gig per card, several cards per server, many servers per rack. You could effectively grep massive ammounts of data in parallel in real time.
Take a look at Zope and/or Plone. Zope is an application server written in python. It is very, very modular and has its own object oriented database built in (ZODB).
Plone is a product on top of Zope that provides something a bit more CMS-like, offering very flexible workflows, custom object types, etc. It is easily 'skinnable' meaning with very little work you can change Plone to look very different (e.g. www.warwickshire.police.uk).
Plone uses a system called Archetypes that allows you to rapidly developed custom content types. There is a produce ArchegenXML that allows you to draw your data model in a UML editor (e.g. ArgoUML, ObjectRealms, etc.) and automatically generate the base code for you.
Some say that Plone is slow, this is relitive, as it is actually doing a lot behind the scenes, and the rapid development more than makes up for this. You can scale Zope/Plone very large -- we are helping a large bank run a Plone-based intranet for 15,000 users. It runs over a cluster of linux and solaris boxes, data is on an EMC storage array and mirrored via fibre to an identical hardware stack on the other side of the city.
"eweek.com reported on the Plone Foundation and CA's involvement, but their information about Plone moving away from the GPL is not correct. This has not happened. For our FAQs on the foundation, please see http://plone.org/foundation/faq.
I'm sitting here with Alan, Paul, and Mark Murphy, and we really want to make sure that every knows that this is a real mistake and we're trying to reach eweek to let them know to issue a retraction. We want to make certain that everyone understands that no changes have been made and that a change like this would never happen with discussion with the community as a whole. The Foundation is an exciting change for our community, and we don't want this mistaken information to let people lose site of that."
Yes, I mean content caching. Zope comes with two cache managers as stanndard (there is an API if you want to writee your own). One is RAMCacheManager, which caches the results of method calls (including page template rendering) to RAM. The other HTTPAcceleratedCacheManager sets HTTP cache headers so an upstream cache (e.g. Apache mod_proxy or Squid) can deal with it.
The caches are very flexible and the RAMCache allows you to cache 'bits' of a page. E.g. we often dynamically build navigation for a site, but RAMCache the results so that the script that traverses the ZODB building the navigation does have to run on every hit.
Zope does have quite high memory requirements. It doesn't need to use all that much to start with (50-60MB) but it has a (tunable) cache that means by default it will use more like 100-200MB on a medium sized site. Our main production Zope servers that host about 60 zope sites runs about 400MB.
There is more to application servers than just raw performance. Hardware nowadays is much cheaper than development time. This is not to say that programmers should be lazy just because they have faster servers. I am just saying that things like dynamic memory allocation and rapid development offer many benefits in term of minimising errors and reducing code maintainence time than raw performance.
Incidentally though, certain parts of Zope, e.g. the security code which is exectuted many times on every request are written in C for added performance.
Zope has many features for tuning speed, including various cache managers.
Err... what has puppies and smiling children got to do with accessibility?
Considering the woman in question is unable to even use the local online phone directory due to poor design, I would consider using a CMS quite an achievement.
Err... there is a *very* active devlopment community for Zope. Just looking at the mailing list archives for last month (March) below are the number of messages posted that month to each list:
Zope 1261 Zope-dev 343 Plone ~1800
That is not to mention the plone-design and other lists, and the #plone and #zope IRC channels of which there are normally on average 5-10 people *actively chatting* on.
Then there is the books published. Admittedly the pace of development has often rendered these books out of date very quickly. However there are at least half a dozen Zope books published, and a Plone book is already available for pre-order at Amazon. I think that is probably more books than any other open source CMS out there currently.
There are also active development sprints being organised around the globe. The most recent being one in DC alongside the python conference last month.
I recently came back from the Plone Sprint in Austria. For those not familiar with sprints, this is where you get a bunch of developers in one place for a week to concentrate on development.
Virtually all of the people there (there were ~50 attendees) ran their own small businesses (myself included, Netsight) that used Plone -- mostly providing installation, customization, and support. Most of these companies *depended* in Plone for their livelihood.
What struck me the most was how business focused all of the developers were. This is something that really sets Plone apart from some of the other OSS projects out there. All of these people are making real dollars on developing this software, and hence *need* to have a business focus otherwise their businesses would fail. As technically great as many OSS projects are, many of them don't have the business drive to succeed.
The second thing that really struck me was a demonstration by a blind woman from the local Institute for the Blind. Plone is known for being very hot on accessibility, but this was just amazing. The woman had half a day training, and was then able to enter content, add metadata and take it through a workflow -- all using a braille reader and text-to-speech software. And what is even more amazing, is that she doesn't speak any English, she was relying on the internationalization features of Plone to deliver a German version of the UI -- including all the alt tags and hidden things that screen-readers rely upon.
Thanks for all the suggestions. It seems that from the replies so far that what I am asking for does not yet exist. Therefore I would like to try and create it is some way.
Specifically: A web site with a database that contains traits of particular mail virii (eg. Subject == 'foobar'; attachment == 'hello.vbs'). A visitor to the site can then download a list of these in such a form that they could copy/paste directly into exim/sendmail/procmail/whatever. The system should be easily extensible such that convertors can be contributed to convert to whatever MTAs people want.
Slashdot Mirror
I had to recently send a backup of data to a client for them to keep for audit purposes and in case we went bust.
I tarred the data up and encypted it with a random string run through aespipe. I then burnt the data to DVD and posted it to them. Once they confirmed they received the DVD I emailed them the key and told them to keep it safe.
Neither the DVD nor the key are any use on their own. The end user doesn't need any PKI setup at all. There is a readme on the DVD telling how to decrypt it, and they are tasked with keeping the key safe.
-Matt
Yes they are going after grep. Very much so. They produce a hardware pattern matcher last time I saw (this was back around 1999). This was a chip on a card with a bunch of DIMMs on it. The theory was several gig per card, several cards per server, many servers per rack. You could effectively grep massive ammounts of data in parallel in real time.
-Matt
Take a look at Zope and/or Plone. Zope is an application server written in python. It is very, very modular and has its own object oriented database built in (ZODB).
Plone is a product on top of Zope that provides something a bit more CMS-like, offering very flexible workflows, custom object types, etc. It is easily 'skinnable' meaning with very little work you can change Plone to look very different (e.g. www.warwickshire.police.uk).
Plone uses a system called Archetypes that allows you to rapidly developed custom content types. There is a produce ArchegenXML that allows you to draw your data model in a UML editor (e.g. ArgoUML, ObjectRealms, etc.) and automatically generate the base code for you.
Some say that Plone is slow, this is relitive, as it is actually doing a lot behind the scenes, and the rapid development more than makes up for this. You can scale Zope/Plone very large -- we are helping a large bank run a Plone-based intranet for 15,000 users. It runs over a cluster of linux and solaris boxes, data is on an EMC storage array and mirrored via fibre to an identical hardware stack on the other side of the city.
-Matt
Joel Burton best sums it up on ZopeZen:
"eweek.com reported on the Plone Foundation and CA's involvement, but their information about Plone moving away from the GPL is not correct. This has not happened. For our FAQs on the foundation, please see http://plone.org/foundation/faq.
I'm sitting here with Alan, Paul, and Mark Murphy, and we really want to make sure that every knows that this is a real mistake and we're trying to reach eweek to let them know to issue a retraction. We want to make certain that everyone understands that no changes have been made and that a change like this would never happen with discussion with the community as a whole. The Foundation is an exciting change for our community, and we don't want this mistaken information to let people lose site of that."
Yes, I mean content caching. Zope comes with two cache managers as stanndard (there is an API if you want to writee your own). One is RAMCacheManager, which caches the results of method calls (including page template rendering) to RAM. The other HTTPAcceleratedCacheManager sets HTTP cache headers so an upstream cache (e.g. Apache mod_proxy or Squid) can deal with it.
The caches are very flexible and the RAMCache allows you to cache 'bits' of a page. E.g. we often dynamically build navigation for a site, but RAMCache the results so that the script that traverses the ZODB building the navigation does have to run on every hit.
Zope does have quite high memory requirements. It doesn't need to use all that much to start with (50-60MB) but it has a (tunable) cache that means by default it will use more like 100-200MB on a medium sized site. Our main production Zope servers that host about 60 zope sites runs about 400MB.
-Matt
There is more to application servers than just raw performance. Hardware nowadays is much cheaper than development time. This is not to say that programmers should be lazy just because they have faster servers. I am just saying that things like dynamic memory allocation and rapid development offer many benefits in term of minimising errors and reducing code maintainence time than raw performance.
Incidentally though, certain parts of Zope, e.g. the security code which is exectuted many times on every request are written in C for added performance.
Zope has many features for tuning speed, including various cache managers.
-Matt
Err... what has puppies and smiling children got to do with accessibility?
Considering the woman in question is unable to even use the local online phone directory due to poor design, I would consider using a CMS quite an achievement.
-Matt
Err... there is a *very* active devlopment community for Zope. Just looking at the mailing list archives for last month (March) below are the number of messages posted that month to each list:
Zope 1261
Zope-dev 343
Plone ~1800
That is not to mention the plone-design and other lists, and the #plone and #zope IRC channels of which there are normally on average 5-10 people *actively chatting* on.
Then there is the books published. Admittedly the pace of development has often rendered these books out of date very quickly. However there are at least half a dozen Zope books published, and a Plone book is already available for pre-order at Amazon. I think that is probably more books than any other open source CMS out there currently.
There are also active development sprints being organised around the globe. The most recent being one in DC alongside the python conference last month.
-Matt
I recently came back from the Plone Sprint in Austria. For those not familiar with sprints, this is where you get a bunch of developers in one place for a week to concentrate on development.
Virtually all of the people there (there were ~50 attendees) ran their own small businesses (myself included, Netsight) that used Plone -- mostly providing installation, customization, and support. Most of these companies *depended* in Plone for their livelihood.
What struck me the most was how business focused all of the developers were. This is something that really sets Plone apart from some of the other OSS projects out there. All of these people are making real dollars on developing this software, and hence *need* to have a business focus otherwise their businesses would fail. As technically great as many OSS projects are, many of them don't have the business drive to succeed.
The second thing that really struck me was a demonstration by a blind woman from the local Institute for the Blind. Plone is known for being very hot on accessibility, but this was just amazing. The woman had half a day training, and was then able to enter content, add metadata and take it through a workflow -- all using a braille reader and text-to-speech software. And what is even more amazing, is that she doesn't speak any English, she was relying on the internationalization features of Plone to deliver a German version of the UI -- including all the alt tags and hidden things that screen-readers rely upon.
--
Matt Hamilton (aka HammerToe)
Netsight Internet Solutions
Where is this on the site? I have heard rumours of this coming, and I've looked on the site, but can't see it anywhere.
-Matt
Thanks for all the suggestions. It seems that from the replies so far that what I am asking for does not yet exist. Therefore I would like to try and create it is some way.
Specifically: A web site with a database that contains traits of particular mail virii (eg. Subject == 'foobar'; attachment == 'hello.vbs'). A visitor to the site can then download a list of these in such a form that they could copy/paste directly into exim/sendmail/procmail/whatever. The system should be easily extensible such that convertors can be contributed to convert to whatever MTAs people want.
Anyone have any suggestions/comments?
-Matt