[ Date Index ] [ Thread Index ] [ <= Previous by date / thread ] [ Next by date / thread => ]
James Fidell wrote: > Tom Potts wrote: > > >> The client and server can exist on the same machine for a 'home' user. >> If you want to be a 'home' user then obviously you would update your >> local server yourself - possibly as part of 'distribution'? >> > > So how does this improve on using M$ Office or OO directly? Surely all > you've actually done is change the applications you have to deal with? > > 1) you don't have to update every machine -just the server 2) you don't have to buy a new operating system when the supplier gives up trying to fix the old one or just want some more cash 3) you don't have to pay for a massive 'server infrastructure' and domain controllers. 4) and most importantly your companies data is accessible. Try finding a document containing a certain phrase on every computer in your 'global enterprise' system now . see below >> If you were part of a company then your data would be stored on a >> central server - as would the programs that run in the browser - they >> are after all only html/css/javascript/xml files. >> The server controls access - you can use directory services if you >> like - which you can control through a browser. >> > > This can already be done using shared filesystems, without the need for > a web browser or server, using existing software (such as M$ Office). > its hard to stop people working locally - because they can they do then see point 4 above. > > If you are one of the > >> awkward squad - i.e. sales on the road then your local copy of the >> software would be updated from the main server when they were online and >> your new data downloaded to the main server at the same time. These are >> transactional operations and can be done safely - databases can >> synchronise very easily these days so that would be the easy way to do >> it. This is only necessary for those people who magically cannot get >> connected when out of the office but still seem to get all the web virii >> going! >> > > And this can already be done with existing systems, though I have to > admit I've no idea how reliable Microsoft's implementation is. Some > very large organisations do it though, presumably quite successfully, > otherwise they wouldn't bother (an ex-client of mine was doing just that > with an installed userbase of somewhere around 7,000 Windows boxes, > but I've no idea how it worked -- I just don't do Windows). > Microsoft software may work but why distribute only to have to recentralise later at vast expense? > >> So home user gets to be god over his own machine and data, a company >> gets to control its own data and not spend a small fortune on installing >> software/os updates and if you do get a salesperson that never connects >> to the internet then you probably win some kind of award. >> > > I don't actually see that you've added much, if anything, to the > functionality that already exists. > massive cost reduction. > > If they lose > >> their laptop - as often seems to happen when called in for a system >> update they can still work on any internet connected machine, and you've >> only lost a limited amount of data if they really have made the effort >> to keep off the web. >> > > Like I'd risk potentially company-confidential data to a random > unknown, unsecured machine (especially in the hands of a salesdroid who > has already demonstrated that they can't be trusted to look after a > laptop :) > they're normally called directors or managers - and they make you do it! > >> IT can be on permanent holiday/on call in the Caymans so long as they >> can get access to the internet somewhere - and don't forget their passwords. >> > > Make that anywhere there's good food/beer/wine, skiing in the winter and > diving in the summer, and I begin to see a compelling argument :) > > A good deal of what you're proposing seems pretty similar to what Sun > (and, IIRC, Oracle) were proposing almost ten years ago when they > started pushing "thin client" technology and the Java workstations. In > many applications their solution was without doubt the right way to go. > Sun's idea was that you'd even download the applications you wanted to > run from a central server to your local workstation, as well as the > data, though everything would be executed by a java runtime rather than > run inside a web browser -- probably a better model if you care about > security. But almost no-one wanted it, and that was well before you > could buy a truckload of PCs with the loose change and fluff you found > down the back of the sofa. I don't see that enough has changed now to > make things any different. > > Before that (and before most people had probably heard of the interweb), > there were things like X terminals -- little boxes with a display and a > copy of the X server in a PROM. Everything you ran had to come from a > centrally-managed server somewhere (and be stored on one, too). Give me > a wifi X terminal now and I could probably do 99.9% of my usual work in > the same way as I already do. Brilliant idea, so where are they now? > (Instead, btw, I'm using a laptop as a glorified X terminal. Before > that I did the same with my desktop PC.) > > The last twenty years (since I've worked in IT) are littered with failed > attempts to centralise management of office-type applications and data > and whilst in some circumstances it makes a great deal of sense and > though a very small number of people have taken the idea on board, it > just doesn't seem to be what lots of people want. I don't think the > will to change that mindset exists yet and I think it's unlikely to > happen in the near future for a number of boring reasons. And when it > happens, if it happens, I doubt it will be via a web browser as we'd > recognise one today. > > James > > It is effectively thin client. The current thinking seems to be - M$ nearly works and we cant be bothered to change peoples mindsets. It may work that way until something forces a change in that mindset - like going out of business to a competitor who saves a fortune on their IT by adopting the KISS approach. The decentralised M$ approach means you pay for a lot of massively complicated software of which only a very tiny proportion of its functionality is used. And you need major systems to run it on. You then have to pay to re-centralise it and then extra for further data mining tools. What percentage of accountants at your firm can actually create a useful pivot table from scratch? What percentage of sales? So why buy all of them a full office suite when you can view then over the web? The current approach is based on what people have been sold - not what is required. A company requires control and access to its data. Your average accounts department can normally find out who agreed to buy printer, who it was bought from, when it was paid for and sometimes even where it is at the click of a mouse. Ask the background for a major corporate decision and its in.. oh who wrote that, no its not there it must be on Harrys laptop, hang on while I try and read this 47 page document... but its really well formatted when you find it! Tom te tom te tom. -- The Mailing List for the Devon & Cornwall LUG http://mailman.dclug.org.uk/listinfo/list FAQ: http://www.dcglug.org.uk/linux_adm/list-faq.html