[Neos] I am sad
Jacob Floyd
cognifloyd at gmail.com
Sun May 18 02:05:19 CEST 2014
Thank you Carsten. We need feedback about what is painful just as much as
we like people to tell us they like Neos. Good luck in your future
projects. I hope you reconsider Neos sooner rather than later.
Cheers,
Jacob Floyd
So, let me try to explain the reasons why i had stop my neos activities.
At first: Reasons are not the amount of bugs and also not missing features.
Its more the experience, that some decisions made under pressure
wich leads to walk into a wrong direction and dont let me trust into the
systems future quality anymore.
Okay, maybe „wrong decisions“ is also wrong. Its more the time when
something ist done.
There are some things, wich had to be done before anything else. If you
dont, you have to rollback
what is currently done.
I want to try to explain this by talking about the following:
1. Caching
2. HTTP
Currently there is NO further work done, either conceptional or practical,
on http handling, wich is the root of any process the Flow Framework will
process.
But whats currently done is conceptional and practical work on caching.
Why am i talking about HTTP here? It seems to work!
But, the concept of http handling is done up to 20% imho.
If you want to make flow realy „Resty“ there are 80% missing.
There is no Patch possible, no Multipart, no image/jpeg content type, etc.
Also there is a mixup done of resource i am requesting and the content of
this resource.
It mixes up GET and POST. Wich is not realy a strict http handling. But
maybe i am wrong.
But if i am right, and you’re developing something else in the layers below
this http layer,
you will run into problems. Working on the http will result in different
structure of arguments, f.e.
This will lead to massive work on routing, mapping, validating AND at least
caching and any request f.e.
wich is done internaly by neos. All of this leads me to 2 more years of
development if you dont rethink your decisions.
If you rethink, it will save you a lot of time on refactoring neos. And it
will also be a fun to develop resty apps.
So looking at the list above, i am sure it needs a switch.
1. HTTP
2. Routing
3. Mapping
4. …
5. …
99999. Caching
The missing HTTP Features i am talking about are known and reported into
the chat.
But i have the feeling, that there is no interest in what i am saying wich
leeds me in:
Losing Trust. And trust ist the most important thing for me. I gave a lot
advance laurels of trust,
but i cant do this anymore.
But let me say again: Its not the thing that there are bugs. I cant trust a
system that have bugs.
But i am loosing trust on wrong conceptions/decisions.
So in my words :) „I give a shit on caching. If its slow, i still able to
do something on the hardware site. But what i definitly need is a perfect
root. If i dont have a perfect root, i cant go further. The risk of
rollback is to high. I dont climb on a tree with the knowledge of a brittle
root.“
I am bound to say that the http handling is wrong in its current state. If
i use it as it is right now, i will end up in unmaintainable projects
because the change will be a breaking one. And all of my work in projects
has to be refactored. My client structure does not give me a chance to do
so. So here i am at a point wich leads me to: Don’t do this!
So this is the root of my decision. But there are more but with less weight.
—————————
I did a try with flow to build up an application wich is multi-client
capable.
For this i do some conceptional work wich leads me to a doctrine feature
called „combined identities“.
Wonderfull, exactly what i need. Doctrine supports it, Flow is using
doctrine … So it should be possible. During development i realized, that
the doctrine implementation in flow seems to be castrated. So, what i tried
is to provide a patch for this. The patch is done for 90%. For the last 10%
i needed help and asked for it. But i dont get it. My feeling: You need to
have a „Standing“ in the community to earn some interest for your patch.
And this is frustrating. Its more frustrating if i see this wake-up call
after my experience: https://www.youtube.com/watch?v=r1Eeb6qO_y0
—————————
Back to Neos
My work with Neos ends up again in this result: Losing Trust
I allways hear that there a live systems and they work well.
What i never here is:
1. „There are live systems, but its very hard to use them because of
cryptical exception messages appearing after 5 klicks“
2. „There are live systems, but we decide only allow access by devs because
they know what they never should do.“
3. „There are live systems, but they are crashing if there are 2 user on it
doing the same work.“
4. „There are live systems, but if you open the backend for users you’ll
end up in a crashed db.“
Are there realy live systems??? I cant believe that they are running
without problems and without anger.
If i will sell Neos to my clients they will send me a lawyer and i can end
up my working.
And i also loose reputation if i recommend to use it. Sorry, but thats the
reality.
But let me say this also: You didnt do a bad job!!!! The idea is good, the
team is highly skilled!!!
But the decissioins seems to have some failures.
Currently i am not in a situation to have the financial background to do
further work with/on neos.
So i decided against it. But i hope you will think about the above said
with focus on conceptional work at the root.
—————————
Its frustrating me if i am at the end of my projekt and decide to cleanup
double media elements from the media module wich ends up in destroying the
work of the last 3 days. deleting a media item in the media browser ends up
in deleting my last 3 days work in creating content because there is no
check of usage of this item in my content. Its deleted hard without any
questions to me „Do you realy want this?“ „HELL, NOOOOOO!“. Ends up in
„Frustration“, end up in „Losing trust“.
Yes, there is a patch for this. But this does not solve another issue. As
far as i remember it simply doing either nothing on the delete request if
its in usage or it throws an exception. But throwing exceptions and dont
handle them isnt a solution.
Wich gets me to the next one:
———————————
The user experience.
My clients having experience on the system. They are calling me by phone
and asking me:
„what exactly means: ‚Unexpected error while moving node:
{„readyState“:4,responseText:“<!Doctype html Public ….“}‘?“
Sure, its hard and complicated to build up a concept how to handle errors
and giving the user not only the problem.
But if its hard, its not a reason to skip it. Instead of this it needs a
Problem/Solution idea or concept. Wich brings me back to the above caching
and a missing/wrong priority list. And i have to say: Building up a cache
or building cloud ready features at this state of work and conception is
waste of time and waste of developer skill. A wrong decision/direction.
Result on my site: Losing trust.
User experience is one of the most important things that neos is able to
earn the recognitions you want for it.
Conception has to be done. Fancy features has to be stopped.
———————————
The „Flow“
You want to get me in a working „Flow“.
But if i try to find answers i find answers decentralized (if i find some)
and not on the typo3.org resources.
I find solutions from sandstorm, kfish etc. The core team has its
infrastructure but there is no central information
about it. Coding guidelines exports for phpstorm, where? Ah, kfish site,
there is an export of php strom settings as far as i can remember. Why is
this all so decentralized. Why are there are so many rare informations and
if there are informations why so hard to find? Thats not a flow you
promised. I bet there are may people who dont know anything about unit
testing. If they start, they ends up in: „Okay, the 2 days of doing stuff
for unit testing was wasting time because flow delivers anything you need“.
Information about this is exactly where? As far as i remember there is no
information of it. but maybe there are some decentralized ones? Maybe i am
wrong and this missing info is fixed now. But there are more. F.e. i am
using xdebug breakpoints. This does not work with flow because of class
compilation. Solution-Search leads me to the sandstorms debug proxy,
decentralized. Trying to use it leads to a waste of time because you
figuere out after some hours, that it does not work anymore with flow.
Infos about this? None, as far as i remember. So my wish: Share your
knowledge, share your infrastructure. And do it centralized. Pick up the
people where they are. Lead them to the promised „Flow“.
———————————
At the end:
Dont get me wrong! I dont criticize any work of any developer! Any
developer in this team has 1000% more skills then me.
My personal fact is „Losing trust“. In most cases you cant realy say why
this happens. But it happens.
Hope you start to ask WHY this happens. You have to do ask, if you want
neos being the big player you want it to be.
I am sure, i am not alone with this feeling.
Kind regards
Carsten
PS I beg your pardon because of my crappy english :)
----------------------
Carsten Bleicker
Wülfingstraße 1
42477 Radevormwald
E-Mail: carsten at bleicker.de
Web: www.bleicker.de
Twitter: @carstenbleicker
Jabber: cbleicker at jabber.ccc.de
Telefon: 02191-7894878
Am 15.05.2014 um 20:32 schrieb Carsten Bleicker <carsten at bleicker.de>:
> Bastian already gets an answer from me. But in german.
> So give me some hours to translate it ;)
>
>
> ----------------------
>
> Carsten Bleicker
> Wülfingstraße 1
> 42477 Radevormwald
>
> E-Mail: carsten at bleicker.de
> Web: www.bleicker.de
> Twitter: @carstenbleicker
> Jabber: cbleicker at jabber.ccc.de
> Telefon: 02191-7894878
>
> Am 15.05.2014 um 20:23 schrieb Rasmus Skjoldan <rasmus at typo3.org>:
>
>> Hi Carsten
>>
>> Just to let you know: I'm listening on here and also very interested in
hearing all the feedback from you.
>>
>> Best Rasmus
>>
>>> On 14/05/2014, at 21.38, Carsten Bleicker <carsten at bleicker.de> wrote:
>>>
>>> I will explain it all tomorrow, okay?
>>>
>>>
>>> ----------------------
>>>
>>> Carsten Bleicker
>>> Wülfingstraße 1
>>> 42477 Radevormwald
>>>
>>> E-Mail: carsten at bleicker.de
>>> Web: www.bleicker.de
>>> Twitter: @carstenbleicker
>>> Jabber: cbleicker at jabber.ccc.de
>>> Telefon: 02191-7894878
>>>
>>>> Am 14.05.2014 um 21:28 schrieb Dominique Feyer <dfeyer at ttree.ch>:
>>>>
>>>> Hi,
>>>>
>>>> I’m really interested too in a definition of « incalculable » and your
personal reason to say that about Neos ?
>>>>
>>>> In two or three year Neos will be really awesome … for a 1.0 projet,
the current situation sound like more than usable.
>>>>
>>>> Bests,
>>>>
>>>> --
>>>> ttree sàrl
>>>> Dominique Feyer
>>>> Rue du Valentin 34 et demi
>>>> CH - 1004 Lausanne
>>>> +41 21 312 36 35
>>>> dfeyer at ttree.ch
>>>> ttree.ch - @ttreeagency - plan d’accès
>>>>
>>>> Le 14 mai 2014 à 19:55:04, Bastian Waidelich (bastian at typo3.org) a
écrit:
>>>>
>>>> Carsten Bleicker wrote:
>>>>
>>>> Hi Carsten,
>>>>
>>>>> after a few months with neos i am realy sorry, but i have to say no
at this point.
>>>>
>>>> That's indeed very sad to read. Your feedback was very valuable.
>>>>
>>>>
>>>>> Imho this system will need another 2 or 3 years to get realy in a
state wich is not only
>>>>> usable but also calculatable for for business usage.
>>>>
>>>> Frankly, Neos is in its infancy still. But for me there's no blocker
>>>> that would make it incalculable (at least not for small and mid-size
>>>> projects). Especially because I feel like being able to adjust behavior
>>>> in a very fine-grained way if I need to.
>>>>
>>>> I'm not trying to convince you of course, but could you elaborate what
>>>> exactly makes it less calculable than CMS for example? Is it the API
>>>> that is still changing a lot, general instability or maybe also the
lack
>>>> of experience you might have with it compared to CMS (no rhetoric
>>>> question!).
>>>>
>>>>
>>>>> Hope to see/read you soon again @all!
>>>>
>>>> Yes! And let's hope that it's not 2 or 3 years from now ;)
>>>>
>>>>
>>>>> Christian, you're so near to my location and i would be happy to
invite you to a lunch for all your help.
>>>>
>>>> In that case, I'm probably close to your location as well and I might
>>>> join you (I'll pay for my lunch of course) ;)
>>>>
>>>> Best,
>>>>
>>>> --
>>>> Bastian Waidelich
>>>> _______________________________________________
>>>> Neos mailing list
>>>> Neos at lists.typo3.org
>>>> http://lists.typo3.org/cgi-bin/mailman/listinfo/neos
>>>> _______________________________________________
>>>> Neos mailing list
>>>> Neos at lists.typo3.org
>>>> http://lists.typo3.org/cgi-bin/mailman/listinfo/neos
>>>
>>> _______________________________________________
>>> Neos mailing list
>>> Neos at lists.typo3.org
>>> http://lists.typo3.org/cgi-bin/mailman/listinfo/neos
>> _______________________________________________
>> Neos mailing list
>> Neos at lists.typo3.org
>> http://lists.typo3.org/cgi-bin/mailman/listinfo/neos
>
> _______________________________________________
> Neos mailing list
> Neos at lists.typo3.org
> http://lists.typo3.org/cgi-bin/mailman/listinfo/neos
_______________________________________________
Neos mailing list
Neos at lists.typo3.org
http://lists.typo3.org/cgi-bin/mailman/listinfo/neos
More information about the Neos
mailing list